JP7091542B1 - Information processing equipment, information processing methods, and information processing programs - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an appropriate service regarding the purchase of receivables.
An information processing device according to the present application has a learning unit and a determination unit. The learning unit learns a learning model that predicts the probability of occurrence of a negative event, which is an event that occurs within a predetermined period by a store opened on a commercial transaction site on the network and is a negative event for the commercial transaction site. The decision unit determines the conditions for purchasing the receivables of the target store based on the probability of occurrence of a negative event by the target store predicted by using the learning model learned by the learning unit.
[Selection diagram] FIG. 8

Description

The present invention relates to an information processing apparatus, an information processing method, and an information processing program.

Conventionally, various techniques related to the purchase of receivables (hereinafter, also referred to as “factoring”) have been provided. For example, a technique for predicting the amount of future receivables and providing funds according to the predicted amount of future receivables is provided.

Japanese Patent No. 6836255 Japanese Patent Application Laid-Open No. 2021-0337662

However, there is room for improvement in the above-mentioned prior art. For example, in the above-mentioned conventional technique, it is difficult to flexibly purchase receivables because the amount of receivables in the future is predicted and the funds are provided according to the amount. Therefore, it is desired to be able to provide an appropriate service regarding the purchase of receivables.

The present application has been made in view of the above, and an object thereof is to provide an information processing device, an information processing method, and an information processing program capable of providing an appropriate service regarding the purchase of a claim.

The information processing device according to the present application learns a learning model for predicting the probability of occurrence of a negative event, which is an event that occurs within a predetermined period by a store opened on a commercial transaction site on the network and is a negative event for the commercial transaction site. A learning unit that determines the conditions for purchasing the claims of the target store based on the probability of occurrence of the negative event by the target store predicted by using the learning model learned by the learning unit. It is characterized by being prepared.

According to one aspect of the embodiment, there is an effect that an appropriate service regarding the purchase of receivables can be provided.

FIG. 1 is a diagram showing an example of information processing according to an embodiment. FIG. 2 is a diagram showing an example of information processing executed by the information processing apparatus. FIG. 3 is a diagram showing an example of information processing executed by the information processing apparatus. FIG. 4 is a diagram showing an example of information processing executed by the information processing apparatus. FIG. 5 is a diagram showing an example of information processing executed by the information processing apparatus. FIG. 6 is a diagram showing an example of information processing executed by the information processing apparatus. FIG. 7 is a diagram showing a configuration example of the information processing system according to the embodiment. FIG. 8 is a diagram showing a configuration example of the information processing apparatus according to the embodiment. FIG. 9 is a diagram showing an example of a store-opener information storage unit according to an embodiment. FIG. 10 is a diagram showing an example of a model information storage unit according to an embodiment. FIG. 11 is a diagram showing an example of the threshold information storage unit according to the embodiment. FIG. 12 is a diagram showing an example of a group information storage unit according to an embodiment. FIG. 13 is a diagram showing an example of the social situation information storage unit according to the embodiment. FIG. 14 is a diagram showing an example of a commercial transaction site information storage unit according to an embodiment. FIG. 15 is a diagram showing an example of a factoring information storage unit according to an embodiment. FIG. 16 is a flowchart showing an example of the flow of information processing executed by the information processing apparatus. FIG. 17 is a flowchart showing an example of the flow of information processing executed by the information processing apparatus. FIG. 18 is a flowchart showing an example of the flow of information processing executed by the information processing apparatus. FIG. 19 is a flowchart showing an example of the flow of information processing executed by the information processing apparatus. FIG. 20 is a flowchart showing an example of the flow of information processing executed by the information processing apparatus. FIG. 21 is a flowchart showing an example of the flow of information processing executed by the information processing apparatus. FIG. 22 is a flowchart showing an example of the flow of information processing executed by the information processing apparatus. FIG. 23 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing device.

Hereinafter, the information processing apparatus, the information processing method, and the embodiment for implementing the information processing program (hereinafter referred to as “the embodiment”) according to the present application will be described in detail with reference to the drawings. It should be noted that this embodiment does not limit the information processing apparatus, information processing method, and information processing program according to the present application. Further, in each of the following embodiments, the same parts are designated by the same reference numerals, and duplicate explanations are omitted.

(Embodiment)
[1. Information processing]
Hereinafter, an example of information processing executed by the information processing system 1 will be described with reference to FIG. FIG. 1 is a diagram showing an example of information processing according to an embodiment. The information processing system 1 executes information processing related to an electronic commerce (EC) service. For example, the information processing apparatus 100 of the information processing system 1 executes information processing regarding the purchase of receivables of a business operator who opens a store on a commercial transaction site on a network (for example, the Internet). The commercial transaction site is, for example, an electronic shopping center (online mall). The commercial transaction site is not limited to the electronic shopping district as long as the information processing shown below can be applied, and may be various forms of commercial transaction sites such as online shopping sites, auction sites, and free market sites. .. Further, in the following, a business operator who opens a store on a commercial transaction site may be described as a store owner, and a business operator who leaves the commercial transaction site after opening a store on the commercial transaction site may be described as a store owner.

In the information processing shown below, a case where future claims are processed is shown as an example of claims. The future claim referred to here is, for example, future sales that have not yet been generated for a store opened on a commercial transaction site. The future bond may be sales that a store opened on the commercial transaction site is expected to generate in the future on the commercial transaction site. The processing for future claims shown below is only an example, and the claims are not limited to future claims and may be any claims as long as the information processing shown below can be applied. For example, the claim may be any claim as long as it is permitted by law to be the target of factoring.

[1-1. Information processing system configuration]
Prior to the description of the processes shown in FIGS. 1 to 6, the configuration of the information processing system 1 will be described with reference to FIGS. 1 and 7. FIG. 7 is a diagram showing a configuration example of the information processing system according to the embodiment. As shown in FIG. 7, the information processing system 1 includes a user terminal 10, a store device 20, a web server 30, and an information processing device 100. The user terminal 10, the store device 20, the web server 30, and the information processing device 100 are connected to each other via a predetermined network N so as to be communicable by wire or wirelessly. Even if the information processing system 1 shown in FIG. 7 includes a plurality of user terminals 10, a plurality of store devices 20, a plurality of web servers 30, and a plurality of information processing devices 100. good.

The user terminal 10 is a computer used by the user. The user terminal 10 is a device (terminal device) that can be carried by the user. For example, the user terminal 10 is used by a user who accesses content such as a web page displayed on a browser or content for an application. For example, the user terminal 10 is used for a user to browse contents (web pages, etc.) of a commercial transaction site, purchase a product on the commercial transaction site, and post an evaluation (so-called review) of product information. To.

The user terminal 10 is realized by, for example, a smartphone, a tablet terminal, a notebook PC (Personal Computer), a mobile phone, a PDA (Personal Digital Assistant), or the like. FIG. 1 shows a case where the user terminal 10 is a smartphone. In the following, the user terminal 10 may be referred to as a user. That is, in the following, the user can be read as the user terminal 10.

The store device 20 is a computer used by a business operator engaged in a business related to the store. For example, the store device 20 is used by a business operator (store owner) of a store (store) that opens a store on a commercial transaction site. The store device 20 transmits / receives information to / from the web server 30. The business operator uses the store device 20 to open a store on the commercial transaction site provided by the web server 30. The store device 20 transmits various information about the store where the business operator opens a store on the commercial transaction site to the web server 30. For example, the store device 20 transmits various information about the store to the information processing device 100. The store device 20 transmits the information requested by the information processing device 100 to the information processing device 100.

The store device 20 is realized by, for example, a smartphone, a tablet terminal, a notebook PC (notebook computer), a desktop PC, a mobile phone, a PDA, or the like. FIG. 1 shows a case where the store device 20 is a notebook computer. In the following, the store device 20 may be referred to as a store or a business operator (store opener) who opened the store. That is, in the following, the store or the business operator (store owner) can be read as the store device 20.

The web server 30 is a server device that provides an electronic commerce service to a user. The web server 30 provides a commercial transaction site such as an electronic shopping street on a network. The web server 30 provides the user with an online shopping service on a commercial transaction site. For example, the web server 30 provides various information about the commercial transaction site in response to a request from the user terminal 10 used by the user, and executes a purchase process of a product that the user wants to purchase.

In the commercial transaction site provided by the web server 30, each business operator opens a store that handles various categories of products. The web server 30 receives a request for opening a store on a commercial transaction site from the store device 20 used by the business operator. The web server 30 receives various information about the store where the business operator opens a store on the commercial transaction site from the store device 20. The web server 30 adds a store in which the requesting business operator opens to the commercial transaction site. Since the information communication between the web server 30 and the store device 20 and the user terminal 10 described above is the same as in the case of online shopping in a general electronic commerce service, detailed description thereof will be omitted.

The web server 30 provides information to the information processing apparatus 100. The web server 30 transmits various information necessary for processing by the information processing apparatus 100 to the information processing apparatus 100. The web server 30 provides information about the commercial transaction site to the information processing apparatus 100. For example, the web server 30 provides various information collected on the commercial transaction site to the information processing apparatus 100. For example, the web server 30 provides the information processing apparatus 100 with various information such as a history of sales of each store on a commercial transaction site and a history of purchases by each user. The web server 30 provides the information collected from the store device 20 and the user terminal 10 to the information processing device 100.

The information processing apparatus 100 is a computer that executes information processing regarding the purchase of receivables of a business operator who opens a store on a commercial transaction site. For example, the information processing device 100 is a determination device that determines conditions (also referred to as “factoring conditions”) relating to the purchase of receivables using various types of information. Further, the information processing apparatus 100 uses various information to determine the upper limit of the purchase of the receivables of the store. Further, the information processing apparatus 100 determines the mode of collecting accounts receivable corresponding to the future receivables of the store based on the estimated (predicted) exit probability of the store. Further, the information processing apparatus 100 determines a proposal destination store to be a proposal destination for the purchase of receivables among a plurality of stores opened on the commercial transaction site.

[1-2. Overview of processing in information processing systems]
From here, an example of information processing performed by the information processing system 1 will be described with reference to FIGS. 1 to 6. First, the outline of the processing in the information processing system 1 including the transmission / reception of information between the devices of the information processing system 1 is described with reference to FIG. 1, and then the details of each information processing are described with reference to FIGS. 2 to 6. explain. The electronic commerce service provided in the information processing system 1 is the same as a general electronic commerce service except for the purchase of receivables, so detailed explanations of points other than the processing related to the purchase of receivables are appropriate. Omit.

First, the store device 20 used by the business operator performs a process of opening the store of the business operator (step S1). For example, the store device 20 requests the web server 30 to open a store according to the operation of the business operator, and the web server 30 adds the business operator's store to the commercial transaction site in response to the request from the store device 20.

Further, the user who uses the user terminal 10 uses the commercial transaction site provided by the web server 30 (step S2). For example, the web server 30 provides the user with an online shopping service on a commercial transaction site in response to a request from the user terminal 10 used by the user. For example, the web server 30 provides various information about the commercial transaction site to the user terminal 10 and executes a purchase process of a product that the user wants to purchase.

Then, the information processing apparatus 100 acquires the information used for processing from the web server 30 (step S3). The information processing apparatus 100 receives information about the commercial transaction site from the web server 30. For example, the information processing apparatus 100 acquires various information collected by the web server 30 from the service of the commercial transaction site from the web server 30. For example, the information processing apparatus 100 acquires various information such as a history of sales of each store on a commercial transaction site and a history of purchases by each user from the web server 30. The information processing device 100 may acquire information from the store device 20 or the user terminal 10.

Then, the information processing apparatus 100 executes various information processing using the acquired information (step S4). The information processing apparatus 100 executes a determination process using the acquired information. For example, the information processing apparatus 100 uses the acquired information to execute a determination process for determining a condition for purchasing a claim, and details of this point will be described later.

Then, the information processing apparatus 100 provides a service to the business operator using the information determined by the determination process (step S5). For example, the information processing apparatus 100 provides an information providing service to a business operator using the information determined by the determination process. The information processing device 100 transmits the information determined by the determination process to the store device 20 used by the business operator. For example, the information processing device 100 transmits information indicating the upper limit of the purchase of the receivables of the store based on the information determined by the determination process to the store device 20 used by the business operator. As described above, the information processing apparatus 100 can provide an appropriate service regarding the purchase of receivables by providing information to the business operator using the information determined by the determination process. The details of the service provided by the information processing apparatus 100 to the store will be described later.

[1-3. Information processing details]
The details of various information processing executed by the information processing apparatus 100 will be described on the premise of the information processing performed by the information processing system 1 described above. The same points as those described above will be omitted as appropriate. In addition, various processes described below may be combined as appropriate.

[1-3-1. Processing based on classification]
For example, the information processing apparatus 100 may perform a determination process using a grouping result in which a plurality of stores are grouped (classified) into a plurality of groups. In this case, the information processing apparatus 100 groups the stores based on the information indicating the probability of being uncollectible calculated based on the store data (store information), and the purchase upper limit month which is an example of the index value for each group. Determine factoring conditions such as numbers. Then, the information processing apparatus 100 determines the purchase upper limit amount by multiplying the purchase upper limit number of months by a value based on the store sales record (the latest one month, etc.). An example of processing based on such classification will be described with reference to FIG. FIG. 2 is a diagram showing an example of information processing executed by the information processing apparatus. In the example of FIG. 2, as an example of the stores opened on the commercial transaction site, only some stores such as stores SP1, SP2, SP12, and SP15 are shown, but many stores are open on the commercial transaction site. It shall be.

In FIG. 2, model M1 is used, which is a learning model (hereinafter, also simply referred to as “model”) that outputs a score (also referred to as “exit score”) indicating the probability (possibility) of the store leaving the commercial transaction site. , The case where the information processing apparatus 100 performs grouping is shown as an example. The model M1 is a store exit model and is a model learned by machine learning. It should be noted that the grouping using the information regarding the closing of the store is only an example, and the information processing apparatus 100 is not limited to the closing of the store, and may perform the grouping of the stores by using the information regarding the risk that the receivables cannot be collected.

For example, the risk of uncollectible receivables includes negative events (also referred to as "negative events") for commercial trading sites. Negative events include various actions (also referred to as "negative actions") depending on the store that can generate the negative events. For example, negative events include violations of the rules, inability to contact, sales of infringing goods, malicious advertisements, counterfeit sales, delinquent store opening fees, plunge in store evaluations, frequent cancellations, delays in delivery, and the like. For example, the information processing apparatus 100 may perform store grouping using information on various risks including the above-mentioned negative events. Further, the information processing apparatus 100 may perform store grouping using not only information on risks such as store closure but also information on the future potential of the store.

Further, the information processing device 100 may acquire the model M1 from a model providing server or the like that provides a learning model, or may learn the model M1 by its own device. For example, when the information processing apparatus 100 learns the model M1, the information processing apparatus 100 learns the model M1 by using the learning data LDT1 stored in the model information storage unit 122 (see FIG. 10). For example, the learning data LDT1 includes data in which information about a store (also referred to as “store information”) and a label indicating whether or not the store has closed (also referred to as “correct answer information”) are associated with each other. ..

The store information includes information on the business of the store operator (business information), information on products traded (sold) in the store (product information), and information on the categories of products traded (sold) in the store (). Various information such as (category information), store sales information (sales information), and store-related evaluation target evaluation information (evaluation information) may be included, but details will be described later. Further, for example, as the correct answer information of each store, "1" is assigned when the store is closed, and "0" is assigned when the store is not closed. For example, as the correct answer information for each store, "1" is assigned to the store information of the store up to a certain time point (time point X) if the store leaves within one year from the time point X, and the store is the time point. If the store has not left within one year from X, "0" may be assigned. In this case, the information processing apparatus 100 estimates (predicts) the probability (possibility) that the store will close within a desired period (for example, one month, six months, five years, etc.), not limited to one year. Can be generated.

For example, the information processing apparatus 100 has backpropagation (error back propagation method) or the like so that the score output by the model M1 approaches the correct answer information (label) associated with the store information of the store input to the model M1. The learning process is performed by the method of. For example, in the information processing apparatus 100, the score output by the model M1 in which the store information including at least one of the store business information, product information, category information, sales information, evaluation information, etc. is input corresponds to the store. The learning process is performed so as to approach the attached correct answer information. For example, the information processing apparatus 100 causes the score output by the model M1 to approach "1" when the correct answer information is "1" and the store information of the store to which the correct answer information is assigned is input. , Perform learning process. Further, for example, in the information processing apparatus 100, when the correct answer information is "0", the score output by the model M1 approaches "0" when the store information of the store to which the correct answer information is assigned is input. As such, the learning process is performed.

For example, the information processing apparatus 100 adjusts the value of the weight (that is, the connection coefficient) considered when the value is transmitted between the nodes by the learning process. In this way, the information processing apparatus 100 learns the model M1 by processing such as backpropagation that corrects parameters (connection coefficients) so that the error between the output in the model M1 and the correct answer information corresponding to the input is reduced. .. For example, the information processing apparatus 100 generates the model M1 by performing processing such as backpropagation so as to minimize a predetermined loss function. As a result, the information processing apparatus 100 can perform a learning process for learning the parameters of the model M1.

The model learning method is not limited to the above-mentioned method, and any known technique can be applied. It should be noted that the generation of each model may be performed by appropriately using various conventional techniques related to machine learning. For example, the model may be generated by using a technique related to machine learning of supervised learning such as SVM (Support Vector Machine). Also, for example, model generation may be performed using techniques related to machine learning for unsupervised learning. For example, model generation may be performed using a technique of deep learning. For example, the model may be generated by appropriately using various deep learning techniques such as DNN (Deep Neural Network), RNN (Recurrent Neural Network), and CNN (Convolutional Neural Network). It should be noted that the description regarding the generation of the above model is an example, and the generation of the model may be performed by a learning method appropriately selected according to the information that can be acquired and the like. That is, if the information processing apparatus 100 can learn the model M1 so as to output the score corresponding to the correct answer information when the store information included in the learning data is input, any method can be used for the model M1. It may be generated.

From here, the information processing shown in FIG. 2 will be described in detail on the premise of the above-mentioned contents. First, in FIG. 2, the information processing apparatus 100 calculates a store exit score for each of a plurality of stores that open on a commercial transaction site. The information processing apparatus 100 inputs the store information SINF1 of the store SP1 into the model M1 (step S11-1). For example, the information processing apparatus 100 inputs the store information SINF1 including at least one of the business information, the product information, the category information, the sales information, the evaluation information, and the like of the store SP1 into the model M1. The model M1 to which the store information SINF1 is input outputs a score (step S11-2). In the example of FIG. 2, the model M1 to which the store information SINF1 is input outputs a score “0.9” as shown in the score SC1. As described above, since the store exit score of the store SP1 is "0.9", which is close to "1", the information processing apparatus 100 can presume that the store SP1 is likely to exit the store from the commercial transaction site. ..

Further, the information processing apparatus 100 inputs the store information SINF2 of the store SP2 into the model M1 (step S11-3). For example, the information processing apparatus 100 inputs the store information SINF2 including at least one of the business information, the product information, the category information, the sales information, the evaluation information, and the like of the store SP2 into the model M1. The model M1 to which the store information SINF2 is input outputs a score (step S11-4). In the example of FIG. 2, the model M1 to which the store information SINF2 is input outputs a score “0.05” as shown in the score SC2. As described above, since the store exit score of the store SP2 is "0.05" and close to "0", it can be estimated that the information processing apparatus 100 is unlikely to leave the store SP2 from the commercial transaction site. ..

Similarly, the information processing apparatus 100 inputs the store information SINF12 of the store SP12 into the model M1 (step S11-5), and the model M1 into which the store information SINF12 is input outputs a score (step S11-6). In the example of FIG. 2, the model M1 to which the store information SINF12 is input outputs a score “0.07” as shown in the score SC12. Further, the information processing apparatus 100 inputs the store information SINF15 of the store SP15 into the model M1 (step S11-7), and the model M1 into which the store information SINF15 is input outputs a score (step S11-8). In the example of FIG. 2, the model M1 to which the store information SINF15 is input outputs the score “0.15” as shown in the score SC15. By the same processing, the information processing apparatus 100 similarly calculates the exit score for stores other than the stores SP1, SP2, SP12, and SP15, but detailed description thereof will be omitted.

Then, the information processing apparatus 100 classifies the plurality of stores into a plurality of groups based on the exit score of each store (step S12). The information processing apparatus 100 groups each store into a plurality of groups by using the exit score of each store and a plurality of threshold values. For example, the information processing apparatus 100 groups each store into a plurality of groups by using the threshold value for grouping stored in the threshold information storage unit 123 (see FIG. 11). In FIG. 2, for the sake of simplicity, a case where nine threshold values increasing at intervals of "0.1" are used will be described as an example. Specifically, the first threshold value TH1 of the value "0.1", the second threshold value TH2 of the value "0.2", the third threshold value TH3 of the value "0.3", and the fourth threshold value of the value "0.4". Threshold TH4, 5th threshold TH5 with value "0.5", 6th threshold TH6 with value "0.6", 7th threshold TH7 with value "0.7", 8th threshold TH8 with value "0.8" , The ninth threshold TH9 with a value of "0.9" is used.

The information processing apparatus 100 groups each store into a plurality of groups according to the comparison result of comparing the exit score of each store with the nine threshold values (first threshold value TH1 to ninth threshold value TH9). As shown in the grouping information GLT1 of FIG. 2, the information processing apparatus 100 groups each store into 10 groups according to the exit score of each store. Specifically, the information processing apparatus 100 includes a first group, a second group, a third group, a fourth group, a fifth group, a sixth group, a seventh group, an eighth group, a ninth group, and a tenth group. Classify each store into one of the groups. The number of groups is not limited to 10, and may be any number such as 5, 15 or the like.

Here, as shown in the grouping information GLT1 of FIG. 2, each group is associated with the number of months for which future claims are to be purchased (also referred to as “number of months to be purchased”). As a result, the information processing apparatus 100 determines the conditions for purchasing the receivables for each group. For example, the information processing apparatus 100 determines that the group to which the store, which is presumed to have a low probability of leaving the store, belongs, the longer the number of months to be purchased is allocated. As described above, the information processing apparatus 100 determines the conditions that are more favorable to the store belonging to the group as the group is less likely to leave the store. That is, the information processing apparatus 100 determines the conditions that are more favorable to the store belonging to the group as the risk is lower.

In the example of FIG. 2, the first group to which the store estimated to have the lowest probability of closing the store belongs is associated with the longest “3” as the number of months to be purchased. In this case, the information processing apparatus 100 indicates that the store belonging to the first group is subject to the purchase of future receivables up to three months ahead.

Further, the shortest "0" as the number of months to be purchased is associated with the tenth group to which the store estimated to have the highest probability of closing the store belongs. In this case, the information processing apparatus 100 indicates that the store belonging to the tenth group will not be purchased in the future. For example, when the information processing apparatus 100 also purchases non-future receivables, the information processing apparatus 100 only purchases non-future receivables for the stores belonging to the 10th group.

For example, the information processing apparatus 100 classifies stores having a store exit score of less than the first threshold value TH1 into the first group. In FIG. 2, the information processing apparatus 100 classifies the store SP2 having a store exit score of “0.05” and having a first threshold value of less than “0.1” into the first group. Further, the information processing apparatus 100 classifies the store SP12 having a store exit score of "0.07" and having a first threshold value of less than "0.1" into the first group.

Further, the information processing apparatus 100 classifies stores having a store exit score of the first threshold value TH1 or more and less than the second threshold value TH2 into the second group. In FIG. 2, the information processing apparatus 100 has a store exit score of “0.15”, is equal to or greater than “0.1” of the first threshold value TH1, and is less than “0.2” of the second threshold value TH2. Is classified into the second group.

Further, the information processing apparatus 100 classifies stores having a store exit score of the second threshold value TH2 or more and less than the third threshold value TH3 into a third group. In FIG. 2, the information processing apparatus 100 has a store exit score (for example, “0.23”) of “0.2” or more of the second threshold value TH2 and less than “0.3” of the third threshold value TH3. Is classified into the third group. Similarly, the information processing apparatus 100 classifies the stores satisfying the conditions into the fourth group to the ninth group by using the third threshold value TH3 to the ninth threshold value TH9.

Further, the information processing apparatus 100 classifies stores having a store exit score of the ninth threshold value TH9 or more into the tenth group. In FIG. 2, the information processing apparatus 100 classifies the store SP1 having a store exit score of “0.9” and having a ninth threshold TH9 of “0.9” or more into the tenth group.

The above-mentioned processing is only an example, and the information processing apparatus 100 may group each store into a plurality of groups by appropriately using various information. For example, the information processing apparatus 100 may be grouped using a value obtained by multiplying the exit score of each store by the sales of each store (also referred to as “damage score”). In this case, the information processing apparatus 100 may be classified into a group in which the condition regarding the purchase of the bond is better as the damage score is lower.

Further, the information processing apparatus 100 may group each store into a plurality of groups without using a threshold value. For example, the information processing apparatus 100 may be grouped based on information on ratios and logarithms. For example, the information processing apparatus 100 may be grouped so that the scores belonging to each group have the same ratio as much as possible. For example, the information processing apparatus 100 may sort each store in order from the one with the lowest exit score, and divide the stores into 10 groups from the first group to the tenth group so that the groups are evenly classified into each group.

Then, the information processing apparatus 100 determines the upper limit of the purchase of future receivables of each store according to the classification result (step S13). For example, the information processing apparatus 100 determines the upper limit of future receivables of each store by using the formula of "upper limit of purchase = sales x number of months to be purchased" as shown in the calculation function FC1 in FIG. Specifically, the information processing apparatus 100 determines the maximum purchase amount of future receivables of the store by multiplying the sales of the store by the number of months to be purchased allocated to the group to which the store belongs. For the sales of a store, the value of sales in the month immediately preceding the store may be used, or the average value of sales in the past month may be used. The average value of the sales of the previous month and the sales of the past month is only an example, and any value is used for the sales of the store as long as it is a value that can be used as the sales of the store for the month. For example, a predicted value (estimated value) of sales in the next month may be used.

For example, when the sales of the store SP1 is the value SL1, the information processing apparatus 100 sets the purchase upper limit to "0 (= SL1 × 0)" for the store SP1 belonging to the tenth group in which the number of months to be purchased is "0". ) ”. The value SL1 is assumed to be a specific numerical value such as "30,000 yen" or "150,000 yen".

Further, when the sales of the store SP2 is the value SL2, the information processing apparatus 100 sets the purchase upper limit to "PL2 (= SL2 × 3)" for the store SP2 belonging to the first group in which the number of months to be purchased is "3". ) ”. The value SL2 is assumed to be a specific numerical value such as "50,000 yen" or "120,000 yen". That is, the purchase upper limit PL2 is three times the sales value SL2. Similarly, the information processing apparatus 100 determines the purchase upper limit for other stores.

The information processing apparatus 100 may adjust the purchase upper limit amount by using other information. For example, the information processing apparatus 100 may increase the purchase upper limit for a store that has a history of using factoring in the past and whose collection has been completed without delay. For example, the information processing apparatus 100 may raise the purchase upper limit for stores that are highly evaluated by users and stores that have high future potential. For example, the information processing apparatus 100 may increase the purchase upper limit when the score output by the future model described later is equal to or higher than a predetermined threshold value.

Then, the information processing apparatus 100 provides information using the determined information (step S14). For example, the information processing apparatus 100 may provide information indicating the purchase upper limit to a store whose purchase upper limit is equal to or greater than a predetermined value. In this case, the information processing device 100 may transmit information indicating the purchase upper limit to the store device 20 of the store whose purchase upper limit is a predetermined value (for example, 100,000 yen or the like) or more. Further, the information processing apparatus 100 may provide information indicating the purchase upper limit amount or the purchase target months to the store in which the purchase target months are equal to or more than a predetermined value. In this case, even if the information processing device 100 transmits information indicating the purchase upper limit amount or the purchase target months to the store device 20 of the store whose purchase target months are a predetermined value (for example, 2 months or the like) or more. good.

In this way, the information processing apparatus 100 classifies the plurality of stores that open on the commercial transaction site into a plurality of groups, and determines the conditions for purchasing the receivables for each of the classified groups. As a result, the information processing apparatus 100 can purchase the receivables according to the determined conditions, so that it is possible to provide an appropriate service regarding the purchase of the receivables. For example, when factoring future receivables, it may be difficult to collect funds if the store is closed in the near future because the future receivables are purchased. Therefore, the information processing apparatus 100 reduces the possibility that it will be difficult to collect funds by grouping according to the risk of store closure and determining index values such as the number of months to be purchased for each group. Can be done.

In the above-mentioned example, the case where the store exit score is calculated using the model M1 which is a learning model is shown as an example, but the information processing apparatus 100 does not use the learning model and the store exit score of each store. May be calculated. In this case, the information processing apparatus 100 may calculate the exit score of each store on a rule basis, for example. The information processing apparatus 100 may calculate the store exit score of each store by using, for example, the rule information in which the condition and the store exit score are associated with each other. The information processing apparatus 100 compares the store information with each rule (condition), and if the store information of the store satisfies the rule, the store exit score of the rule may be used as the store exit score of the store. Further, the information processing apparatus 100 may compare the store information with the rules (conditions) belonging to each group, and classify the store into a group in which the store information of the store satisfies the rule.

[1-3-2. Processing based on the model corresponding to the period]
The processing based on the above-mentioned grouping is only an example, and the information processing apparatus 100 may determine the conditions for purchasing the receivables without performing the grouping. In this case, the information processing apparatus 100 uses the store data (store information) to generate a learning model that estimates the probability (possibility) of a negative event such as a store exit or a violation of the rules within a certain period. Based on the estimated information, determine the factoring conditions such as the maximum number of months for future receivables to be purchased.

In FIG. 3, the closing of the store from the commercial transaction site will be described as an example of a negative event. , Malicious advertising, counterfeit sales, store opening fee delinquency, sudden drop in store evaluation, frequent cancellations, delivery delays, etc. For example, the information processing apparatus 100 estimates a period in which the possibility that a store will be closed meets a predetermined criterion by using a plurality of models that output a score indicating the probability (possibility) of the store closing for each period. , Determine factoring conditions based on that period.

This point will be described with reference to FIG. FIG. 3 is a diagram showing an example of information processing executed by the information processing apparatus. In FIG. 3, a process based on a model corresponding to a period will be described as an example. In the example of FIG. 3, the case where the store SP12 is the target store for determining the factoring condition (also referred to as “target store”) will be described as an example, but the target store is not limited to the store SP12, and any of the stores will be opened on the commercial transaction site. It may be a store of. The same points as described above will be omitted as appropriate.

In FIG. 3, models M11, M12, M13, M14, etc., which are learning models (models) that output a score (store exit score) indicating the probability (possibility) of the store leaving the store within a predetermined period from the commercial transaction site, are shown. An example is shown in which the information processing apparatus 100 performs determination processing using a plurality of models. Models M11, M12, M13, M14, etc. are store exit models corresponding to different periods, and are models learned by machine learning.

Here, the model M11 is a model that outputs a score indicating the probability (possibility) that the target store leaves the commercial transaction site within one month from that point. The model M12 is a model that outputs a score indicating the probability (possibility) that the target store will leave the commercial transaction site within two months from that point. The model M13 is a model that outputs a score indicating the probability (possibility) that the target store will leave the commercial transaction site within three months from that point. The model M14 is a model that outputs a score indicating the probability (possibility) that the target store will leave the commercial transaction site within four months from that point. Although only models M11, M12, M13, and M14 are shown in FIG. 3, models M15 and six months that output a score indicating the probability (possibility) that the target store will leave the commercial transaction site within five months. A model corresponding to each period such as a model M16 that outputs a score indicating the probability (possibility) that the target store leaves the commercial transaction site may be used during the period. Further, the period is not limited to one month, and may be any interval such as one week.

Further, the information processing apparatus 100 may acquire a plurality of models M11, M12, M13, M14, etc. from a model providing server or the like that provides a learning model, or may acquire a plurality of models M11, M12, M13, M14 by its own device. Etc. may be learned. For example, when the information processing apparatus 100 learns the model M11, the information processing apparatus 100 learns the model M11 by using the learning data LDT11 stored in the model information storage unit 122 (see FIG. 10). For example, in the learning data LDT11, whether or not the store has closed within one month from the time when the information about the store (store information) and the latest information in the store information is acquired (also referred to as "starting time"). The data associated with the label (correct answer information) indicating the above is included.

For example, as the correct answer information for each store, "1" is assigned if the store is closed within one month from the start, and "1" is assigned if the store is not closed within one month from the start. "0" is assigned. The information processing apparatus 100 performs learning processing so that the score output by the model M11 into which the store information of the store is input approaches the correct answer information associated with the store. For example, the information processing apparatus 100 causes the score output by the model M11 to approach "1" when the correct answer information is "1" and the store information of the store to which the correct answer information is assigned is input. , Perform learning process. Further, for example, in the information processing apparatus 100, when the correct answer information is "0", the score output by the model M11 approaches "0" when the store information of the store to which the correct answer information is assigned is input. As such, the learning process is performed. Since the learning process of the model M11 is the same as the learning process of the model M1, detailed description thereof will be omitted.

Further, when learning the model M12, the information processing apparatus 100 learns the model M12 by using the learning data LDT12 stored in the model information storage unit 122. For example, the learning data LDT12 contains information about the store (store information) and a label indicating whether or not the store has closed within two months from the time when the latest information in the store information was acquired (at the time of calculation). The data associated with the correct answer information) is included.

For example, as the correct answer information for each store, "1" is assigned if the store is closed within two months from the start, and "1" is assigned if the store is not closed within two months from the start. "0" is assigned. The information processing apparatus 100 performs learning processing so that the score output by the model M12 into which the store information of the store is input approaches the correct answer information associated with the store. For example, the information processing apparatus 100 causes the score output by the model M12 to approach "1" when the correct answer information is "1" and the store information of the store to which the correct answer information is assigned is input. , Perform learning process. Further, for example, in the information processing apparatus 100, when the correct answer information is "0", the score output by the model M12 approaches "0" when the store information of the store to which the correct answer information is assigned is input. As such, the learning process is performed. Since the learning process of the model M12 is the same as the learning process of the models M1 and M11, detailed description thereof will be omitted.

Similarly, when learning the model M13, the information processing apparatus 100 learns the model M13 using the learning data LDT13 stored in the model information storage unit 122, and when learning the model M14, the model information storage unit 122. The model M14 is trained using the training data LDT14 stored in. For example, the learning data LDT13 contains information about the store (store information) and a label indicating whether or not the store has closed within three months from the time when the latest information in the store information was acquired (at the time of calculation). The data associated with the correct answer information) is included. For example, the learning data LDT14 contains information about the store (store information) and a label indicating whether or not the store has closed within four months from the time when the latest information in the store information was acquired (at the time of calculation). The data associated with the correct answer information) is included. Since the learning process of the models M13 and M14 is the same as the learning process of the models M1, M11 and M12, detailed description thereof will be omitted. If the information processing apparatus 100 can learn the models M11, M12, M13, and M14 so as to output the score corresponding to the correct answer information when the store information included in the learning data is input, what if the information processing apparatus 100 can learn? Models M11, M12, M13, and M14 may be generated by any method.

From here, the information processing shown in FIG. 3 will be described in detail on the premise of the above-mentioned contents. First, in FIG. 3, the information processing apparatus 100 calculates a store exit score for each period of the store SP12, which is the target store. The information processing apparatus 100 inputs the store information SINF12-1 of the store SP12 into the model M11 that predicts the probability (possibility) of leaving the store in one month (step S21-1). For example, the information processing apparatus 100 inputs the store information SINF12-1 including at least one of the business information, the product information, the category information, the sales information, the evaluation information, and the like of the store SP12 into the model M11. The model M11 to which the store information SINF12-1 is input outputs the store exit score (step S21-2). In the example of FIG. 3, the model M11 in which the store information SINF12-1 is input has a store exit score “0.02” indicating the probability (possibility) of exiting the store in one month as shown in the score SC12-1. Is output. As described above, since the store exit score indicating the probability (possibility) of store exit during one month of the store SP12 is "0.02", which is close to "0", the information processing apparatus 100 has the store SP12. It can be estimated that it is unlikely that the store will be closed from the commercial transaction site within a month from that point.

Further, the information processing apparatus 100 inputs the store information SINF12-2 of the store SP12 into the model M12 that predicts the probability (possibility) of leaving the store within two months (step S21-3). For example, the store information SINF12-2 is the same store information as the store information SINF12-1. When the input of the model is different for each period, the store information SINF12-2 may be different from the store information SINF12-1, and may be store information having different collection periods, for example. The model M12 in which the store information SINF12-2 is input outputs the store exit score (step S21-4). In the example of FIG. 3, the model M12 in which the store information SINF12-2 is input has a store exit score “0.2” indicating the probability (possibility) of exiting the store within two months as shown in the score SC12-2. Is output.

Similarly, the information processing apparatus 100 inputs the store information SINF12-3 of the store SP12 into the model M13 that predicts the probability (possibility) of leaving the store within three months (step S21-5), and stores the store information SINF12. The model M13 in which -3 is input outputs the exit score (step S21-6). For example, the store information SINF12-3 is the same store information as the store information SINF12-1. When the input of the model is different for each period, the store information SINF12-3 may be different from the store information SINF12-1 and the store information SINF12-2. In the example of FIG. 3, the model M13 in which the store information SINF12-3 is input has a store exit score “0.52” indicating the probability (possibility) of exiting the store within three months as shown in the score SC12-3. Is output.

Further, the information processing apparatus 100 inputs the store information SINF12-4 of the store SP12 into the model M14 for predicting the probability (possibility) of leaving the store within four months (step S21-7), and the store information SINF12- The model M14 in which 4 is input outputs the exit score (step S21-8). In the example of FIG. 3, the model M14 in which the store information SINF12-4 is input has a store exit score “0.65” indicating the probability (possibility) of exiting the store within four months as shown in the score SC12-4. Is output. In this way, the information processing apparatus 100 calculates a store exit score whose value increases as the period becomes longer. Even if the information processing apparatus 100 calculates a store exit score indicating the probability (possibility) of leaving the store within 5 months and a store exit score indicating the probability (possibility) of leaving the store within 6 months. good.

Then, the information processing apparatus 100 determines the number of months for the future receivables of the store SP12 to be purchased based on the exit score of the store SP12 for each period (step S22). For example, the information processing apparatus 100 estimates the time when the store SP12 is likely to leave the store by using the threshold value for estimating the store closing time stored in the threshold information storage unit 123 (see FIG. 11). In FIG. 3, a case where the value of the threshold value TH11 regarding the estimation of the closing time is 0.5 will be described as an example.

The information processing apparatus 100 compares the store exit score output by each model with the threshold value TH11, and determines the number of months to purchase future receivables of the target store based on the comparison result. For example, the information processing apparatus 100 determines the period corresponding to the learning model having the shortest period among the learning models whose exit score is the threshold value TH11 or more as the number of months for the purchase of future receivables of the target store. In FIG. 3, the information processing apparatus 100 compares the store exit score output by the models M11, M12, M13, M14, etc. with the threshold value TH11 value “0.5”, and is the target month for the purchase of future receivables of the store SP12. Determine the number. Since the store exit score "0.52" output by the model M13 that predicts the probability (possibility) of leaving the store in three months is equal to or higher than the threshold value TH11, the information processing apparatus 100 has the target number of months information in FIG. As shown in TM1, the number of months for the purchase of future receivables of the store SP12 is determined to be "3".

It should be noted that the above-mentioned processing is only an example, and the information processing apparatus 100 may determine the number of months for future purchase of future receivables of the target store by appropriately using various information. For example, the information processing apparatus 100 sets the period shorter than the period corresponding to the learning model having the shortest period among the learning models having the exit score of the threshold TH11 or more as the number of months for the purchase of future bonds of the target store. You may decide. In this case, the information processing apparatus 100 sets the period (for example, 2 months) shorter than the period (3 months) corresponding to the model M13 in which the exit score is the threshold value TH11 or more to the purchase target month of the future receivables of the store SP12. You may decide on a number. Further, for example, the information processing apparatus 100 may determine the number of months for future purchase of receivables by using the value obtained by accumulating the exit scores of each period. For example, the information processing apparatus 100 calculates the cumulative score by accumulating the exit scores in ascending order of months, compares the cumulative score with the threshold value, and purchases the number of months when the cumulative score exceeds the threshold value. You may decide on the number of months. In this case, the information processing apparatus 100 has a cumulative score of "0.22" when, for example, the exit score after one month is "0.02" and the exit score after two months is "0.2". (= 0.02 + 0.2) ”. When the threshold value is "0.2", the information processing apparatus 100 determines the number of months "2" for which the cumulative score "0.22" exceeds the threshold value "0.2" as the number of months to be purchased.

Then, the information processing apparatus 100 determines the upper limit of the purchase of future receivables of the store SP12 based on the determined number of months to be purchased and the sales information of the store SP12, because it is the same as the process described with reference to FIG. Detailed explanation will be omitted. For example, the information processing apparatus 100 uses the calculation function FC1 (see FIG. 2) to determine the purchase upper limit of the store SP12 as "PL2 (= SL2 × 3)" when the sales of the store SP2 is the value SL2. ..

Then, the information processing apparatus 100 provides information using the determined information (step S23). Since the information provision is the same as the process described with reference to FIG. 2, detailed description will be omitted. However, for example, the information processing apparatus 100 provides information indicating to the business operator of the store SP12 the number of months to be purchased or the maximum purchase amount. May be provided. In this case, the information processing device 100 may transmit information indicating the number of months to be purchased or the maximum purchase amount of the store SP12 to the store device 20 used by the business operator of the store SP12.

In this way, the information processing apparatus 100 predicts the store exit score for each period of the target store, and determines the conditions for purchasing the receivables using the predicted store exit score for each period. As a result, the information processing apparatus 100 can purchase the receivables according to the determined conditions, so that it is possible to provide an appropriate service regarding the purchase of the receivables. For example, when factoring future receivables, it may be difficult to collect funds if the store is closed in the near future because the future receivables are purchased. Therefore, the information processing apparatus 100 predicts a period in which the target store is likely to close, and determines an index value such as the number of months to be purchased according to the period, so that it may be difficult to collect funds. The sex can be reduced.

As described above, for example, the information processing apparatus 100 can increase the amount of receivables to be purchased by estimating the period until the store is closed and setting the sales in that period as the purchase target period. The above-mentioned processing is only an example, and the information processing apparatus 100 determines factoring conditions using various information. For example, the information processing apparatus 100 may use a plurality of models for estimating the probability of occurrence of each of the negative events. In this case, the information processing apparatus 100 may estimate the store exit probability using a plurality of scores output by the plurality of models corresponding to the negative event. Even if the information processing apparatus 100 calculates the occurrence probability of each of the negative events, estimates the store exit probability based on the calculated occurrence probability, and determines the factoring condition based on the estimated store exit probability. good.

[1-3-3. Processing based on social conditions]
The information processing apparatus 100 is not limited to the store information, and various information may be used. For example, the information processing apparatus 100 may use a social situation such as a social economic situation. In this case, the information processing apparatus 100 estimates the total number of closed stores (hereinafter, also simply referred to as "number of closed stores" or "number of closed stores") based on social conditions such as stock prices, and uses the estimated number of closed stores. Purchase conditions may be set based on this. For example, the information processing apparatus 100 reflects the estimated number of closed stores in the grouping of stores.

In FIG. 4, the number of stores closing from the commercial transaction site (number of closed stores) will be described as an example of the number of occurrences of negative events, but as described above, the number of negative events to be estimated is the negative event. Not limited to the closing of the store, it may be a violation of the rules, inability to contact, sale of infringing goods, malicious advertisement, fake sale, delinquent store opening fee, sudden drop in store evaluation, frequent cancellations, delay in delivery, etc. As described above, the number of target negative events is not limited to the number of store closures, and may be the number of occurrences of various negative events. For example, the information processing apparatus 100 estimates the number of closed stores using social situation information indicating the economic situation of society, and determines the factoring condition based on the estimated number of closed stores.

This point will be described with reference to FIG. FIG. 4 is a diagram showing an example of information processing executed by the information processing apparatus. FIG. 4 shows an example of processing based on social conditions. In FIG. 4, a process of changing the threshold value used for grouping based on the number of store closures estimated based on the social situation will be described as an example. Specifically, FIG. 4 describes, as an example, a process of changing the threshold value used in FIG. 2 based on the number of store closures estimated based on the social situation. The same points as described above will be omitted as appropriate.

FIG. 4 shows, as an example, a case where the information processing apparatus 100 changes the threshold value by using the model M31 which is a learning model (model) that outputs a score indicating the number of closed stores corresponding to the social situation. The model M31 is a model used for estimating the number of closed stores and is a model learned by machine learning. In FIG. 4, a case where the model M31 outputs the number of closed stores as a score indicating the number of closed stores will be described as an example, but the score output by the model M31 is not limited to the number of closed stores itself, but the number of closed stores. It may be a value used for calculation (for example, a ratio to the number of stores closed at that time).

Further, the information processing device 100 may acquire the model M31 from a model providing server or the like that provides a learning model, or may learn the model M31 by its own device. For example, when the information processing apparatus 100 learns the model M31, the information processing apparatus 100 learns the model M31 by using the learning data LDT31 stored in the model information storage unit 122 (see FIG. 10). For example, in the learning data LDT31, data in which information on a social situation (also referred to as "social situation information") and a label indicating the number of store closures corresponding to the social situation (also referred to as "correct answer information") are associated with each other are stored. included.

The social situation information includes information indicating various social situations such as information on the economy (economic situation). For example, social situation information includes various information indicating economic conditions such as stock prices, unemployment rate, and interest rates. For example, as the correct answer information of the social situation, the number of closed stores at the time corresponding to the social situation is associated. FIG. 4 shows, as an example, a case where the number of stores closed three months after the time point X is assigned to the social situation information at a certain time point (time point X) as the correct answer information of each social situation. The information processing apparatus 100 generates a model M31 that estimates the number of store closures six months after the time corresponding to the social situation. The information processing device 100 estimates the number of stores that will be closed in the desired future (for example, one month, six months, one year, five years, etc.) corresponding to the time of the social situation, not limited to three months later. It is possible to generate a model to be (predicted).

For example, the information processing apparatus 100 performs learning processing so that the score output by the model M31 approaches the correct answer information (label) associated with the social situation information of the store input to the model M31. For example, in the information processing apparatus 100, when the correct answer information is "100", the score output by the model M31 approaches "100" when the social situation information of the store to which the correct answer information is assigned is input. In addition, the learning process is performed. Further, for example, in the information processing apparatus 100, when the correct answer information is "5", the score output by the model M31 becomes "5" when the social situation information of the store to which the correct answer information is assigned is input. The learning process is performed so as to approach. When a store is closed from a commercial transaction site, it becomes impossible to collect the receivables from the store. Therefore, in FIG. 4, the information processing apparatus 100 estimates the number of store closures from the commercial transaction site as the number of uncollectible receivables. Learn the model M31 used to do this. Since the learning process of the model M31 is the same as the learning process of the model M1 and the like, detailed description thereof will be omitted. If the information processing apparatus 100 can learn the model M31 so as to output the score corresponding to the correct answer information when the social situation information included in the learning data is input, any method can be used for the model M31. May be generated.

From here, the information processing shown in FIG. 4 will be described in detail on the premise of the above-mentioned contents. First, the information processing apparatus 100 inputs the social situation information SDT 11 indicating the social situation at the time of processing (current) into the model M31 (step S31-1). For example, the information processing apparatus 100 inputs the social situation information SDT 11 including at least one such as the current stock price, the unemployment rate, and the interest rate into the model M31. The model M31 to which the social situation information SDT11 is input outputs a score (step S31-2). In the example of FIG. 4, the model M31 in which the social situation information SDT11 is input outputs the number of store closures “NW1” as shown in the score SC31. The number of store closures NW1 is assumed to be a specific value (for example, 200 or the like) indicating the estimated number of store closures. As described above, in FIG. 4, the information processing apparatus 100 estimates the number of store closures from the commercial transaction site as the number of uncollectible receivables.

Then, the information processing apparatus 100 determines the threshold value based on the estimated number of store closures (step S32). In FIG. 4, the information processing apparatus 100 changes the threshold information TL1 used up to that point to the threshold information TL2. For example, the information processing apparatus 100 lowers the threshold value when the estimated number of store closures NW1 satisfies a predetermined condition. For example, the information processing apparatus 100 lowers the threshold value when the number of store closures NW1 is larger than the number of store closures at that time (currently) (also referred to as “currently measured value”). In FIG. 4, for the sake of simplicity, a case where the information processing apparatus 100 lowers each threshold value by 0.5 will be described as an example, but the information processing apparatus 100 has the number of store closures at that time (current). The threshold value may be changed by increasing or decreasing the threshold value according to the comparison between (currently measured value) and the number of closed stores NW1. In this case, the information processing apparatus 100 may determine the threshold value on the condition that the number of exits NW1 is different from the measured value at the present time.

For example, the information processing apparatus 100 may determine the threshold value according to the comparison result between the current measured value and the number of stores closed NW1. For example, the information processing apparatus 100 may lower each threshold value according to the difference or ratio between the current measured value and the number of stores closed NW1. For example, when the number of store exits NW1 with respect to the current measured value is "0.9", the information processing apparatus 100 may update each threshold value by multiplying each threshold value by "0.9". For example, when the number of store exits NW1 with respect to the current measured value is "1.05", the information processing apparatus 100 may update each threshold value by multiplying each threshold value by "1.05". Further, the information processing apparatus 100 may change each threshold value individually. For example, the information processing apparatus 100 may lower only the first threshold value TH1 according to the difference or ratio between the number of stores closed and the number of stores closed NW1 at that time (currently). The above is only an example, and the information processing apparatus 100 may change the threshold value by processing using various information as appropriate.

Then, the information processing apparatus 100 classifies the plurality of stores into a plurality of groups by using the changed threshold information TL2 (step S33). Since the processing after step S33 is the same as the processing after step S12 in FIG. 1 except for the value of each threshold value, detailed description thereof will be omitted.

In this way, the information processing apparatus 100 changes the threshold value based on the number of store closures estimated based on the economic situation, classifies the plurality of stores into a plurality of groups using the changed threshold value, and classifies the group. For each, the conditions for purchasing the receivables are determined. As a result, the information processing apparatus 100 can perform processing using a threshold value appropriately adjusted according to the economic situation and purchase the receivables, so that it is possible to provide an appropriate service regarding the purchase of the receivables. Can be done. For example, when factoring future receivables, it may be difficult to collect funds if the store is closed in the near future because the future receivables are purchased. Therefore, the information processing apparatus 100 can reduce the possibility that it will be difficult to collect funds by performing grouping in consideration of the economic situation and determining index values such as the number of months to be purchased for each group. can.

In the above-mentioned example, the case of calculating the number of store closures using the model M31, which is a learning model, is shown as an example, but the information processing apparatus 100 calculates the number of store closures without using the learning model. You may. In this case, the information processing apparatus 100 may calculate the number of store closures, for example, on a rule basis. The information processing apparatus 100 may calculate the number of store closures by using, for example, rule information in which a condition and the number of store closures are associated with each other. The information processing apparatus 100 compares the social situation information with each rule (condition), and if the social situation information satisfies the rule, the number of store closures of the rule may be used. Further, the information processing apparatus 100 may determine the conditions for purchasing the receivables without grouping. In this case, the information processing apparatus 100 may determine the number of months for future purchase of receivables according to the estimated number of store closures (number of occurrences). For example, the information processing apparatus 100 uses a function in which the estimated number of store closures (number of occurrences) is a variable (explanatory variable) and the number of months to be purchased is an output value (objective variable), and the number of months to be purchased in the future. May be determined.

As described above, the information processing apparatus 100 estimates the number of store closures based on the social situation (economic situation), and changes the threshold used for grouping based on the estimated number of store closures. The information processing apparatus 100 estimates the number of stores to be closed in the future based on the economic situation, groups them using a threshold changed based on the estimated number of stores to be closed, and determines the factoring condition. In this way, the information processing apparatus 100 can determine the factoring condition of the store based on the estimated number of store closures by grouping the stores into groups associated with the factoring condition.

The information processing apparatus 100 may estimate the number of store closures based on the predicted information indicating the future social situation (economic situation), and change the threshold value used for grouping based on the estimated number of store closures. .. In this case, the information processing apparatus 100 may acquire forecast information indicating a future social situation (economic situation) provided from a predetermined external device, and may estimate the number of store closures using the acquired forecast information. .. Further, the information processing apparatus 100 may estimate the number of store closures based on the social situation (economic situation) and change the factoring conditions of each group based on the estimated number of store closures. For example, the information processing apparatus 100 may reduce the number of months to be purchased by each group when the estimated number of store closures is a predetermined number or more. Further, for example, the information processing apparatus 100 may increase the number of months to be purchased by each group when the estimated number of store closures is less than a predetermined number.

The above-mentioned processing is only an example, and the information processing apparatus 100 may perform the processing using various information. For example, the information processing apparatus 100 may determine the conditions for purchasing the receivables based on the number of closed stores at that time (currently) and the estimated number of closed stores. For example, the information processing apparatus 100 may determine the factoring condition based on the relative numerical value of the number of store closures. For example, the information processing apparatus 100 sets the number of closed stores at that time (current) and the present as "1", and when the estimated number of closed stores is "1.2", which is 1.2 times that number, factoring. The condition may be reduced by 20%. In this case, when the current purchase target month of one store is "2", the information processing apparatus 100 may reduce the purchase target month of the one store to "1.6".

Further, for example, in the information processing device 100, when the current total number of closed stores is 100, the economic situation is on the decline, and the number of closed stores is estimated to be 200 next year, the threshold value for grouping is changed. May be good. In this case, for example, the information processing apparatus 100 may change the value of a certain threshold value from "0.6" to "0.5". As a result, the information processing apparatus 100 may change the store exit score of 0.6-0.8 belonging to a certain group to 0.5-0.8. The information processing apparatus 100 may reduce the number of months to be purchased.

Further, the information processing apparatus 100 may predict the number of store exit scores for each of various factors. For example, the information processing apparatus 100 may predict the number of closed stores for each product handled on the commercial transaction site. In this case, the information processing apparatus 100 may predict the number of stores to be closed for each product based on the economic situation of each product. For example, when the information processing apparatus 100 is a product (trend product) that is popular at that time, the number of store exit scores may be predicted to be low for that product. For example, if the home-based goods are prevalent at the time when an infectious disease such as a coronavirus is prevalent, the information processing apparatus 100 may predict that the number of exit scores for products related to the home-based goods is low.

[1-3-4. Processing based on category]
The information processing apparatus 100 may determine the factoring conditions of the store according to the category of the products handled by the store. In this case, the information processing apparatus 100 determines the factoring conditions for the receivables of the target store based on the past sales trends of the target store and the past sales trends of each category of the products handled on the commercial transaction site. May be good.

In FIG. 5, a category (also referred to as “main category”) corresponding to the products (main products) mainly handled by the store and a category (“sub-category”) corresponding to the products (sub-products) handled by the store other than the main products are used. The case where the processing is performed according to (also referred to as) is shown as an example. This point will be described with reference to FIG. FIG. 5 is a diagram showing an example of information processing executed by the information processing apparatus. In FIG. 5, the processing based on the category will be described as an example. If the information processing apparatus 100 determines the factoring conditions for the receivables of the target store according to the category of the products handled by the target store, the receivables of the target store can be stored without classifying the categories such as the main category and the subcategories. Factoring conditions may be determined.

In the example of FIG. 5, the case where the store SP15 is the target store (target store) for determining the factoring condition will be described as an example, but the target store is not limited to the store SP15, and any store opened on the commercial transaction site. You may. The same points as described above will be omitted as appropriate.

FIG. 5 shows, as an example, a case where the information processing apparatus 100 performs a determination process using a model M41 which is a learning model (model) that outputs a score for estimating a sales tendency for each category. The model M41 is a category sales model that estimates the sales tendency of the store corresponding to the category, and is a model learned by machine learning. Although FIG. 5 shows the case where the model M41 common to the main category and the sub-category is used, individual models may be used for the main category and the sub-category.

Further, the information processing apparatus 100 may acquire the model M41 or the like from a model providing server or the like that provides a learning model, or may learn the model M41 or the like by its own device. For example, when the information processing apparatus 100 learns the model M41, the information processing apparatus 100 learns the model M41 by using the learning data LDT41 stored in the model information storage unit 122 (see FIG. 10). For example, in the learning data LDT41, information on one category of products handled by one store (category information by store), information on the one category on a commercial transaction site (category information in the site), and one in the one store. Contains data associated with labels (correct answer information) indicating sales trends in the category of.

For example, the learning data LDT41 includes store-specific category information including information indicating sales of products belonging to one category in one store at a certain time point (time point X), and the entire commercial transaction site at a certain point in time (time point X). The in-site category information including the information indicating the sales of the products belonging to the one category in the above is associated with the label (correct answer information) indicating the change after the time point X of the sales of the one category in the one store. Contains data. The store-specific category information may include information indicating changes in sales of products belonging to one category in the store up to a certain point in time (time point X). For example, the store-specific category information may include information on a plurality of sales of products belonging to one category in a store corresponding to a plurality of time points up to a certain time point (time point X).

The in-site category information may include information indicating changes in sales of products belonging to one category in the entire commercial transaction site up to a certain point in time (time point X). For example, the in-site category information may include information on a plurality of sales of products belonging to one category in the entire commercial transaction site corresponding to a plurality of time points up to a certain time point (time point X). In addition, the category information for each store and the category information in the site may be described as model input information together. For example, a value indicating the relationship between the sales of one category in one store at the time point X and the time point (time point Y) after a predetermined period has elapsed from the time point X is associated with the correct answer information.

For example, if the sales of one category in one store at time point Y is 1.5 times the sales of one category in one store at time point X (that is, sales increase), the value "1.5" is correct. Associated as information. In this way, when the sales of one category in one store increase after the time point X, a value larger than 1 is associated as correct answer information. If the sales of one category in one store at time point Y is 0.6 times the sales of one category in one store at time point X (that is, the sales decrease), the value "0.6" is the correct answer information. Can be associated. For example, when the sales of one category in one store decrease after the time point X, a value smaller than 1 is associated as correct answer information. That is, the learning data LDT 41 includes information indicating whether or not one category is an uptrend or a downtrend from a certain time point (time point X) as correct answer information. The above is only an example, and various information may be used for the correct answer information. For example, the value of sales at time point Y itself may be used for the correct answer information. In this case, the model M41 outputs a score indicating an estimated value of sales of the category in the target store.

The information processing apparatus 100 performs learning processing so that the score output by the model M41 in which the model input information corresponding to one category of the target store is input approaches the correct answer information associated with the model input information. For example, in the information processing apparatus 100, when the correct answer information is "1.2", the score output by the model M41 becomes "1.2" when the model input information to which the correct answer information is assigned is input. The learning process is performed so as to approach. Further, in the information processing apparatus 100, when the correct answer information is "0.8" and the model input information to which the correct answer information is assigned is input, the score output by the model M41 is "0.8". The learning process is performed so as to approach. Since the learning process of the model M41 is the same as the learning process of the model M1 and the like, detailed description thereof will be omitted. Further, if the information processing apparatus 100 can learn the model M41 so as to output the score corresponding to the correct answer information when the model input information included in the learning data is input, any method can be used for the model M41. May be generated.

From here, the information processing shown in FIG. 5 will be described in detail on the premise of the above-mentioned contents. First, the information processing apparatus 100 determines the main category and the sub-category of the store SP15. For example, the information processing apparatus 100 refers to the store information of the store SP15 and determines the main category and the sub-category of the store SP15. For example, the information processing apparatus 100 may determine the category with the highest sales among the categories handled by the store SP15 as the main category. For example, the information processing apparatus 100 may determine the category having the largest number of products among the categories handled by the store SP15 as the main category. Further, the information processing apparatus 100 may determine a subcategory other than the main category among the categories handled by the store SP15.

The determination of the main category and the sub-category described above is only an example, and the information processing apparatus 100 may determine the main category and the sub-category by appropriately using various information. The information processing apparatus 100 may determine a category having the next highest sales after the main category among the categories handled by the store SP15 as a sub-category. Further, the information processing apparatus 100 may determine a category having the second largest number of products next to the main category as a subcategory among the categories handled by the store SP15.

The information processing apparatus 100 may determine the main category and the sub-category of the store SP15 by using the category information stored in the store information storage unit 121 (see FIG. 9). When the main category and the sub-category of the store SP15 are specified by the business operator of the store SP15, the information processing apparatus 100 determines the main category and the sub-category of the store SP15 according to the designation by the business operator of the store SP15. .. In the following, a case where the main category of the store SP15 is determined to be category # 1 and the subcategory is determined to be category # 2 will be described as an example. Category # 1 and category # 2 are, for example, specific categories such as food, beverage, liquor, clothing, and cosmetics.

The information processing apparatus 100 inputs the model input information MINF15-1 corresponding to the category # 1, which is the main category of the store SP15 at the time of processing (current), into the model M41 (step S41-1). For example, the information processing apparatus 100 indicates store-specific category information including information indicating sales of products belonging to category # 1 of the current store SP15, and sales of products belonging to that category # 1 in the entire current commercial transaction site. The model input information MINF15-1 including the in-site category information including the information is input to the model M41. The model M41 to which the model input information MINF15-1 is input outputs a score (step S41-2). In the example of FIG. 5, the model M41 to which the model input information MINF15-1 is input outputs the score “VL41” as shown in the score SC31. The main score VL41 is assumed to be a value (for example, 0.4, 2.5, etc.) indicating a trend of an estimated trend for the main category (category # 1) of the store SP15.

Further, the information processing apparatus 100 inputs the model input information MINF15-2 corresponding to the category # 2, which is a subcategory of the store SP15 at the time of processing (current), into the model M41 (step S41-3). For example, the information processing apparatus 100 indicates store-specific category information including information indicating sales of products belonging to category # 2 of the current store SP15, and sales of products belonging to that category # 2 in the entire current commercial transaction site. The model input information MINF15-2 including the in-site category information including the information is input to the model M41. The model M41 to which the model input information MINF15-2 is input outputs a score (step S41-4). In the example of FIG. 5, the model M41 to which the model input information MINF15-2 is input outputs the score “VL42” as shown in the score SC31. The subscore VL42 is assumed to be a value (for example, 0.6, 1.8, etc.) indicating a trend of the estimated trend for the subcategory (category # 2) of the store SP15.

Then, the information processing apparatus 100 determines the upper limit of the purchase of future receivables of the store SP15 based on the trend trend of the product category handled by the store SP15 (step S42). In FIG. 5, the information processing apparatus 100 determines the number of months for future receivables of the store SP15 to be purchased based on the trend trend of the product category handled by the store SP15. For example, the information processing apparatus 100 uses a function (“number of months”) in which the score of the main category (main score) and the score of the subcategory (subscore) are variables (explanatory variables) and the number of months to be purchased is an output value (objective variable). The number of months for the purchase of future receivables of the store SP15 is determined by using the calculation function FC2).

For example, the number of months calculation function FC2 is purchased by the first value obtained by multiplying the main score by the first coefficient (first weight) and the second value obtained by multiplying the subscore by the second coefficient (second weight). It is a function to calculate the number of months. For example, the number of months calculation function FC2 outputs the total value of the first value and the second value as the number of months to be purchased. For example, the information processing apparatus 100 determines the value of each weight so that the first weight corresponding to the score (main score) of the main category is larger than the second weight corresponding to the score (subscore) of the subcategory. You may. Further, the information processing apparatus 100 may calculate the second value by multiplying the subscore by a value obtained by multiplying the second weight by a predetermined attenuation value (for example, 0.8 or the like).

In FIG. 5, the information processing apparatus 100 is a purchase target of the store SP15 by the first value obtained by multiplying the main score VL41 of the store SP15 by the first weight and the second value obtained by multiplying the subscore VL42 of the store SP15 by the second weight. The value VL43 indicating the number of months is calculated. The value VL43 is a specific value (for example, 1.5, 3, etc.) indicating the number of months to be purchased.

Here, the information processing apparatus 100 may determine the weighting of the category by using various information. The information processing apparatus 100 may determine the value of the first weight and the value of the second weight using various information. For example, the information processing apparatus 100 may be weighted by the number of products in the category or by the sales in the category. For example, the information processing apparatus 100 may increase the weight value as the number of products in the category increases, or may increase the weight value as the number of sales in the category increases. For example, the information processing apparatus 100 may increase the value of the first weight as the sales in the main category increase. For example, the information processing apparatus 100 may increase the value of the second weight as the number of products in the subcategory increases.

Then, the information processing apparatus 100 determines the upper limit of the purchase of future receivables of the store SP15 based on the determined number of months to be purchased and the sales information of the store SP15, because it is the same as the process described with reference to FIG. Detailed explanation will be omitted. The information processing apparatus 100 determines the purchase upper limit amount of the store SP15 as "PL41" as shown in the purchase upper limit information PL1 of FIG. For example, the information processing apparatus 100 uses the calculation function FC1 (see FIG. 2) to determine that the purchase upper limit of the store SP15 is “PL41 (= SL15 × VL43)” when the sales of the store SP15 is the value SL15. ..

Then, the information processing apparatus 100 provides information using the determined information (step S43). Since the information provision is the same as the process described with reference to FIG. 2, detailed description will be omitted. However, for example, the information processing apparatus 100 provides information indicating to the business operator of the store SP15 the number of months to be purchased or the maximum purchase amount. May be provided. In this case, the information processing device 100 may transmit information indicating the number of months to be purchased or the maximum purchase amount of the store SP15 to the store device 20 used by the business operator of the store SP15.

In this way, the information processing apparatus 100 determines the conditions for purchasing the receivables by using the information of the category for the target store. As a result, the information processing apparatus 100 can purchase the receivables according to the category of the products handled by the target store, so that it is possible to provide an appropriate service regarding the purchase of the receivables. For example, when factoring future receivables, it may be difficult to collect funds if the store is closed in the near future because the future receivables are purchased. Therefore, the information processing apparatus 100 can reduce the possibility that it will be difficult to collect funds by determining an index value such as the number of months to be purchased according to the trend of the category of the product handled by the target store.

In the above-mentioned example, the information processing apparatus 100 calculates a score indicating an uptrend and a downtrend for each category of the products handled by the store, and determines the factoring condition. The above-mentioned processing is only an example, and the information processing apparatus 100 may determine factoring conditions using various information. For example, the information processing apparatus 100 may raise the purchase upper limit if the score in the main category is high. Further, the information processing apparatus 100 may have a larger weight as the category has a larger number of products. Further, in the information processing apparatus 100, the weighting may be increased as the sales amount increases. As described above, the main category and the like may be registered in advance for each store. Further, when the target store handles a category that is popular as a whole, the information processing apparatus 100 may determine the factoring condition of the target store based on the share (occupancy rate) in the category.

Further, the information processing apparatus 100 may change the weighting depending on whether or not the category with the largest number of products handled in the target store is the main category. For example, the information processing apparatus 100 may determine the factoring condition of the target store depending on whether a predetermined condition is satisfied, such as a tendency of a specific category to rise or fall more prominently than usual. Further, for example, even if the sales of the information processing device 100 are on a downward trend, if the information processing device 100 is a store that handles a popular category (for example, corona-related goods) at that time, the information processing device 100 of the store The factoring conditions may be improved.

[1-3-5. Processing based on a model from multiple perspectives]
The information processing apparatus 100 is not limited to a model from one viewpoint, and may perform processing based on a model from a plurality of viewpoints. In this case, the information processing apparatus 100 has a store exit model (store exit prediction model) that predicts the store exit from the commercial transaction site and a future potential model (future possibility prediction model) that predicts the future potential of the store. Factoring conditions such as the maximum number of months for future loan purchases are determined using these two models.

In FIG. 6, the closing of the store from the commercial transaction site will be described as an example of a negative event. , Malicious advertising, counterfeit sales, store opening fee delinquency, sudden drop in store evaluation, frequent cancellations, delivery delays, etc. For example, the information processing apparatus 100 uses two different viewpoint models, a model M1 which is a store closing model and a model M2 which is a future model, to determine the probability (possibility) of closing a store for each period. Using multiple models that output the indicated scores, the conditions for purchasing the store's receivables are determined.

This point will be described with reference to FIG. FIG. 6 is a diagram showing an example of information processing executed by the information processing apparatus. FIG. 6 describes an example of processing based on a model from a plurality of viewpoints. In the example of FIG. 5, the case where the store SP12 is the target store (target store) for determining the factoring condition will be described as an example, but the target store is not limited to the store SP12, and any store opened on the commercial transaction site. You may. The same points as described above will be omitted as appropriate. For example, since the model M1 used in FIG. 6 is the same as the model M1 in FIG. 2, the description of the learning process and the like will be omitted.

In FIG. 6, the information processing apparatus 100 is determined by using the model M2, which is a learning model (model) that outputs a score (also referred to as “future potential score”) indicating the future potential of the store opened on the commercial transaction site. An example of processing is shown. The model M2 is a future model and is a model learned by machine learning.

Further, the information processing device 100 may acquire the model M2 from a model providing server or the like that provides a learning model, or may learn the model M2 by its own device. For example, when the information processing apparatus 100 learns the model M2, the information processing apparatus 100 learns the model M2 by using the learning data LDT2 stored in the model information storage unit 122 (see FIG. 10). For example, the learning data LDT2 contains information about the store (store information) and a label indicating how the sales of the store have changed since the latest information in the store information was acquired (acquisition time). Data associated with (correct answer information) is included.

For example, as the correct answer information for a store, "1.5" is assigned if the store's sales increase 1.5 times (that is, sales increase) after a predetermined period has elapsed (for example, one month) from the time of acquisition. If the sales increase 0.8 times (that is, the sales decrease) after a predetermined period has elapsed from the time of acquisition of the store, "0.8" is assigned. The information processing apparatus 100 performs learning processing so that the score output by the model M2 into which the store information of the store is input approaches the correct answer information associated with the store. For example, in the information processing apparatus 100, when the correct answer information is "1.8", the score output by the model M2 is "1.8" when the store information of the store to which the correct answer information is assigned is input. The learning process is performed so as to approach. Further, for example, in the information processing apparatus 100, when the correct answer information is "0.2", the score output by the model M2 is "0." when the store information of the store to which the correct answer information is assigned is input. The learning process is performed so as to approach "2". In this case, a model M2 is generated that outputs a score greater than 1 if the store has a high future in terms of sales, and outputs a score less than 1 if the store has a high future in terms of sales. .. Since the learning process of the model M2 is the same as the learning process of the model M1, detailed description thereof will be omitted.

From here, the information processing shown in FIG. 6 will be described in detail on the premise of the above-mentioned contents. First, in FIG. 6, the information processing apparatus 100 calculates a store exit score and a future potential score for the store SP12, which is the target store. The information processing apparatus 100 inputs the store information SINF12 of the store SP12 at the time of processing (current) into the model M1 (step S51-1). The model M1 to which the store information SINF12 is input outputs a score (step S51-2). In the example of FIG. 6, the model M1 to which the store information SINF12 is input outputs the score “VL51” as shown in the score SC51. It is assumed that the store exit score VL51 is a value (for example, 0.1, 0.75, etc.) indicating the probability (possibility) that the store SP12 will exit the store from the commercial transaction site.

Further, the information processing apparatus 100 inputs the store information SINF12 of the store SP12 at the time of processing (current) into the model M2 (step S51-3). The model M2 in which the store information SINF12 is input outputs a score (step S51-4). In the example of FIG. 6, the model M2 in which the store information SINF12 is input outputs the score “VL52” as shown in the score SC52. The future potential score VL52 is assumed to be a value (for example, 0.6, 1.5, etc.) indicating the future potential of the sales of the store SP12. In FIG. 6, for the sake of simplicity, the case where the input information of the model M1 and the model M2 are the same is shown, but the input information of the model M1 and the model M2 may be different.

Then, the information processing apparatus 100 determines the upper limit of the purchase of future receivables of the store SP12 based on the store exit score VL51 and the future potential score VL52 of the store SP12 (step S52). In FIG. 6, the information processing apparatus 100 determines the number of months for the purchase of future receivables of the store SP12 based on the store exit score VL51 and the future potential score VL52 of the store SP12. For example, the information processing apparatus 100 uses a function (also referred to as "month number calculation function FC3") in which the store exit score and the future potential score are variables (explanatory variables) and the number of months to be purchased is an output value (objective variable). Then, the number of months for the purchase of future receivables of the store SP12 is determined.

For example, the month number calculation function FC3 has a first value obtained by multiplying the reciprocal of the exit score by the first coefficient (first weight), and a second value obtained by multiplying the future potential score by the second coefficient (second weight). It is a function to calculate the number of months to be purchased based on the value. For example, the month number calculation function FC3 outputs the total value of the first value and the second value as the number of months to be purchased.

In FIG. 6, the information processing apparatus 100 has a first value obtained by multiplying the reciprocal of the exit score VL51 of the store SP12 by the first weight, and a second value obtained by multiplying the future potential score VL52 of the store SP12 by the second weight. The value VL53 indicating the number of months to be purchased of the store SP12 is calculated. The value VL53 is a specific value (for example, 1.5, 3, etc.) indicating the number of months to be purchased.

The value of the first weight and the value of the second weight may be set as appropriate. For example, when the risk of leaving a store is important, the value of the first weight may be set large. Further, for example, when the future potential of the store is important, the value of the second weight may be set large.

Then, the information processing apparatus 100 determines the upper limit of the purchase of future receivables of the store SP12 based on the determined number of months to be purchased and the sales information of the store SP12, because it is the same as the process described with reference to FIG. Detailed explanation will be omitted. As shown in the purchase upper limit information PL2 of FIG. 6, the information processing apparatus 100 determines the purchase upper limit amount of the store SP12 as “PL51”. For example, the information processing apparatus 100 uses the calculation function FC1 (see FIG. 2) to determine the purchase upper limit of the store SP12 as “PL51 (= SL12 × VL53)” when the sales of the store SP12 is the value SL12. ..

Then, the information processing apparatus 100 provides information using the determined information (step S53). Since the information provision is the same as the process described with reference to FIG. 2, detailed description will be omitted. However, for example, the information processing apparatus 100 provides information indicating to the business operator of the store SP12 the number of months to be purchased or the maximum purchase amount. May be provided. In this case, the information processing device 100 may transmit information indicating the number of months to be purchased or the maximum purchase amount of the store SP12 to the store device 20 used by the business operator of the store SP12.

In this way, the information processing apparatus 100 determines the conditions for purchasing the receivables of the target store from a plurality of viewpoints. As a result, the information processing apparatus 100 can purchase the receivables of the target store in consideration of a plurality of viewpoints comprehensively, so that it is possible to provide an appropriate service regarding the purchase of the receivables. For example, when factoring future receivables, it may be difficult to collect funds if the store is closed in the near future because the future receivables are purchased. Therefore, the information processing apparatus 100 can reduce the possibility that it will be difficult to collect funds by determining index values such as the number of months to be purchased that comprehensively considers the risk of store closure and the future potential of the target store. can.

As described above, the information processing apparatus 100 can improve the processing accuracy while suppressing the enormous amount of learning data by dividing the store exit model and the future potential model. The above-mentioned processing is only an example, and the information processing apparatus 100 may perform the processing using various information. For example, the store exit model may be a model trained to input past store data and output whether or not the store has closed within a predetermined period. Further, the future potential model is not limited to whether or not sales will increase in the future, and may be a model that predicts the number of visitors, the number of products, the number of store evaluations, and the like. For example, the information processing apparatus 100 may determine the factoring conditions of the store depending on whether or not the handling ratio of the popular (buzzing) products is high.

[1-3-6. Determination of recovery mode]
Next, an example of determining the mode of collecting money when factoring is performed will be described. For example, the information processing apparatus 100 determines a mode of collecting accounts receivable corresponding to future receivables when factoring regarding future receivables to the store is performed. The information processing apparatus 100 determines the lower limit of the collection rate based on the store closing probability. For example, the information processing apparatus 100 determines the lower limit of the monthly redemption rate according to the probability of leaving the store after a predetermined period. For example, the information processing apparatus 100 predicts the store exit probability every month and determines the lower limit of the collection rate so that the collection is completed by the month when the store exit probability exceeds a predetermined threshold value.

For example, the information processing apparatus 100 calculates a store exit score for each period, determines a period in which the store exit score exceeds a threshold value as a collection period, and determines a lower limit so that the entire amount can be collected within the determined collection period. In this case, the information processing apparatus 100 determines the recovery period using a plurality of models, each of which corresponds to a different period as shown in FIG. This point will be described with reference to FIG. In the following, a case where the store SP12 is the target store and the store SP12 performs factoring regarding future receivables will be described as an example. Further, the same points as the above-mentioned processing will be omitted as appropriate.

For example, the information processing apparatus 100 determines a mode of collecting accounts receivable corresponding to future receivables of the store SP12 based on a store exit score indicating the store exit probability of the store SP12 for each period. For example, the information processing apparatus 100 determines the collection period of accounts receivable corresponding to the future receivables of the store SP12 by using the threshold value for estimating the closing time stored in the threshold information storage unit 123 (see FIG. 11).

The information processing apparatus 100 compares the store exit score output by each model with the threshold value TH11, and determines the collection period of accounts receivable corresponding to the future receivables of the target store based on the comparison result. For example, the information processing apparatus 100 determines the period corresponding to the learning model having the shortest period among the learning models having the exit score of the threshold TH11 or more as the collection period of accounts receivable corresponding to the future receivables of the target store. .. The information processing apparatus 100 compares the exit score output by the models M11, M12, M13, M14, etc. with the threshold value TH11 value "0.5" to determine the collection period of accounts receivable corresponding to the future receivables of the store SP12. decide. Since the store exit score "0.52" output by the model M13 that predicts the probability (possibility) of store exit within three months is equal to or higher than the threshold value TH11, the information processing apparatus 100 corresponds to future receivables of the store SP12. Set the collection period of accounts receivable to "three months".

The above-mentioned processing is only an example, and the information processing apparatus 100 may appropriately use various information to determine the collection period of accounts receivable corresponding to future receivables. For example, the information processing apparatus 100 determines a period shorter than the period corresponding to the learning model having the shortest period among the learning models having a store exit score of the threshold TH11 or more as the collection period of accounts receivable corresponding to future receivables. You may. In this case, the information processing apparatus 100 has accounts receivable corresponding to future receivables of the store SP12 for a period shorter than the period (three months) corresponding to the model M13 in which the exit score is equal to or higher than the threshold value TH11 (for example, two months). May be determined during the payback period.

Then, the information processing apparatus 100 determines the collection rate of the accounts receivable corresponding to the future receivables of the store SP12 according to the determined collection period. For example, the information processing apparatus 100 determines the collection rate so that the entire amount can be collected within the determined collection period. For example, if the information processing apparatus 100 determines the collection period of accounts receivable corresponding to future receivables of the store SP12 to be three months, the monthly payment rate (collection rate) of the store SP12 so that the entire amount can be collected within three months. Is determined to be "33.4%".

Then, the information processing apparatus 100 notifies the target store of the determined information. For example, the information processing device 100 has a collection period of "three months" and a monthly payment rate (collection rate) of "33.4%" in the store device 20 used by the business operator of the store SP12. Information indicating that is sent.

In this way, the information processing apparatus 100 predicts the exit score for each period of the target store, and determines the collection mode of the accounts receivable corresponding to the future receivables of the target store according to the predicted period. As a result, the information processing apparatus 100 can determine the collection mode in consideration of the risk that the target store will be closed, so that it is possible to provide an appropriate service regarding the purchase of receivables. For example, when factoring future receivables, it may be difficult to collect funds if the store is closed in the near future because the future receivables are purchased. Therefore, the information processing apparatus 100 predicts a period in which the target store is likely to close, and determines the collection period and the collection rate according to the period, thereby reducing the possibility that the collection of funds will be difficult. Can be made to.

By the process described above, the information processing apparatus 100 can determine the lower limit of the recovery rate so that the recovery period does not extend even if the recovery rate can be set by the store side. As described above, the information processing apparatus 100 determines the lower limit of the collection rate of the receivables. For example, the information processing apparatus 100 may determine the collection mode so that 10% of the factored amount is collected every month when the probability of leaving the store becomes maximum after 10 months. For example, the information processing apparatus 100 may determine the collection rate so that the entire amount can be collected in nine months when the probability of leaving the store becomes maximum after nine months.

If the collection of accounts receivable corresponding to the future receivables of the target store satisfies a predetermined condition, the information processing apparatus 100 may update the collection mode of the accounts receivable corresponding to the future receivables of the target store. For example, when the collection of accounts receivable corresponding to the future receivables of the target store is delayed, the information processing apparatus 100 updates the closing probability of the target store, and the future of the target store is based on the updated closing probability of the target store. Update the collection mode of accounts receivable corresponding to receivables. For example, when the period exceeding the threshold value is changed from 5 months to 4 months, the information processing apparatus 100 may shorten the recovery period by one month and determine the recovery rate in that period.

As described above, the information processing apparatus 100 may extend the collection period or recalculate the collection rate when the store delays payment. Further, the information processing apparatus 100 may recalculate the recovery rate when refactoring.

[1-3-7. Service to the store]
From here, an example of providing a service to the store will be described. The information processing apparatus 100 provides a service to the store in step S5 shown in FIG. The information processing apparatus 100 determines a store (also referred to as a “proposal destination store”) to which factoring is proposed. For example, the information processing apparatus 100 determines, among the plurality of stores, the proposed store to which the purchase of the receivable is proposed, based on the store information regarding each of the plurality of stores opened on the commercial transaction site.

Then, the information processing apparatus 100 provides information for proposing the purchase of the bond to the determined proposal destination store. For example, the information processing apparatus 100 determines the number of months to be purchased and the maximum purchase amount of the determined proposed store by the determination process as shown in FIGS. 2 to 5, and the determined number of months to be purchased. And provide information indicating the maximum purchase amount to the proposed store. For example, the information processing device 100 is a store device in which the business operator of the proposed store uses information indicating the number of months to be purchased or the maximum purchase amount of the proposed store determined by the determination process as shown in FIGS. 2 to 5. Send to 20.

Here, the information processing apparatus 100 may determine the proposal destination store using various information. For example, the information processing apparatus 100 may determine a proposal destination store by using store information including information indicating whether or not the store is advertising. For example, the information processing apparatus 100 may determine a proposal destination store depending on whether or not an advertisement is being placed. For example, the information processing apparatus 100 determines a store that does not advertise as a proposal destination store. As a result, the information processing apparatus 100 can propose factoring to a store that is not currently advertising and is expected to increase sales if it is advertised.

Further, for example, the information processing apparatus 100 may determine the proposal destination store by using the store information including the information indicating the handled product which is the product handled by the store. For example, the information processing apparatus 100 may determine the proposal destination store according to the products handled by the store. For example, the information processing apparatus 100 determines a store in which the number of products handled is less than a predetermined threshold value as the proposed store. As a result, the information processing apparatus 100 can propose factoring to a store where the number of product items is small and sales are expected to increase if the number of product items is increased.

Further, the information processing apparatus 100 learns the characteristics of the relationship between the store opened on the commercial transaction site and the change in sales when the store takes measures for increasing sales (also referred to as "execution measures"). A model (also referred to as a “measure effect prediction model”) may be used to determine the store to which the proposal is made. In this case, the information processing apparatus 100 determines as the proposal destination store a store whose score output by the proposal destination determination model is equal to or higher than a predetermined threshold value. In this case, the information processing apparatus 100 proposes an execution measure to the proposed store together with the factoring conditions. Further, the information processing apparatus 100 may propose a factoring condition to the proposal destination store on condition that the execution measure is executed.

For example, the measure effect prediction model is a model in which the store information of a store and the measure information indicating the execution measures taken by the store are input, and the value indicating the increase or decrease in sales when the store takes measures is output. Is. For example, in the information processing apparatus 100, model input information including store information of a store and measure information indicating execution measures taken by the store, and a value (correct answer information) indicating an increase or decrease in sales corresponding to the model input information are used. The measure effect prediction model is learned using the associated learning data. The information processing apparatus 100 may acquire a measure effect prediction model from a model providing server or the like that provides a learning model.

For example, the information processing apparatus 100 may determine a factoring proposal destination by using a model (measure effect prediction model) that learns a change (increase / decrease) in sales when a measure such as an advertisement is taken. When the information processing apparatus 100 outputs a score indicating that sales increase when the measure effect prediction model implements a measure such as an advertisement for a certain store, the information processing apparatus 100 may determine the store as the proposal destination store.

As described above, the information processing apparatus 100 determines the factoring proposal destination store. The above-mentioned processing is only an example, and the information processing apparatus 100 may perform the processing by appropriately using various information. For example, if the information processing apparatus 100 has a store that can be extended by changing the place where the advertisement is placed, factoring may be proposed to the store. Further, for example, if the information processing apparatus 100 handles products in a category with future potential and a store that grows if the place where the advertisement is placed is changed, factoring may be proposed to the store. Further, the information processing apparatus 100 may propose information indicating a user (target user) to whom the advertisement is provided and factoring conditions when the advertisement is performed to the target user.

[2. Information processing device configuration]
Next, the configuration of the information processing apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 8 is a diagram showing a configuration example of the information processing apparatus 100 according to the embodiment. As shown in FIG. 8, the information processing apparatus 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The information processing device 100 includes an input unit (for example, a keyboard, a mouse, etc.) that receives various operations from the administrator of the information processing device 100, and a display unit (for example, a liquid crystal display, etc.) for displaying various information. You may have.

(Communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card) or the like. Then, the communication unit 110 is connected to the network by wire or wirelessly, and transmits / receives information to / from the user terminal 10, the store device 20, and the web server 30.

(Memory unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory element such as a RAM (Random Access Memory) or a flash memory (Flash Memory), or a storage device such as a hard disk or an optical disk. As shown in FIG. 8, the storage unit 120 according to the embodiment includes a store store information storage unit 121, a model information storage unit 122, a threshold information storage unit 123, a group information storage unit 124, and a social situation information storage unit. It has 125, a commercial transaction site information storage unit 126, and a factoring information storage unit 127. Further, the storage unit 120 is not limited to the above, and may store various information.

(Store owner information storage unit 121)
The store-opener information storage unit 121 according to the embodiment stores information about a business operator that opens a store on a commercial transaction site. For example, the store-opener information storage unit 121 stores various information about the store (store) in which the business operator (store-opener) has opened a store on the commercial transaction site. For example, the information stored in the store clerk information storage unit 121 is acquired from the web server 30 via the communication network. FIG. 9 is a diagram showing an example of a store-opener information storage unit according to an embodiment. The store-opener information storage unit 121 shown in FIG. 9 displays items such as "store-opener", "store", "store-opening date", "closing date", "reason for leaving", "store information", and "score information". Have. Further, the "store information" includes items such as "business information", "product information", "category information", "sales information", and "evaluation information". In addition, the "score information" includes items such as "closed" and "future potential".

The "store owner" indicates information (business ID or name, etc.) that identifies a business operator that opens a store on a commercial transaction site. For example, "store owner" indicates information for identifying a business operator who has opened a store in the past or is currently opening a store on a commercial transaction site. "Store" indicates a store (store) in which a business operator has opened a store on a commercial transaction site.

Although the store-opener information storage unit 121 shown in FIG. 9 shows an example in which each business operator opens one store on the commercial transaction site, each business operator may open a plurality of stores on the commercial transaction site. In this case, the business operator can use the store properly according to the category of the products to be handled and the business form. For example, a process may be performed in which another store of a business operator having a store closed due to forced closure or the like is also closed (the right to open the store is lost).

"Store opening date" indicates the date when the store was opened on the commercial transaction site. "Store closing date" indicates the date when the store was closed from the commercial transaction site. For businesses (or stores) that have not closed at this time, the item for the closing date will be blank (“-” (hyphen)).

"Store information" includes various information about the store. The "store information" includes information registered by the store operator (registration data) and history (actual data) accumulated by the operation of the store. For example, the "store information" includes information used by the information processing apparatus 100 for learning the model.

"Business information" indicates information about a business. "Business information" includes various information related to the business such as information indicating the business form of a corporation or sole proprietor, information on business funds such as capital, information indicating the type of business of the store, and the number of years the store has continued. In addition, the "business information" may include information about the business operator. For example, the "business information" may include attribute information of the business operator and the like. In addition, the "business information" may include financial information indicating the financial status of the business operator. The above is only an example, and the "business information" may include various information about the business. In the example of FIG. 9, although it is indicated by an abstract code such as “business information # 1”, the business information corresponding to the store is stored.

"Product information" indicates information about products traded in the store. "Product information" includes various information related to products such as product identification information (JAN code or product ID, etc.), product name, product image, product description, product price, product inventory quantity, product category, etc. Contains information. The above is only an example, and the "product information" may include various information about the product. In the example of FIG. 9, although it is indicated by an abstract code such as “product information # 1”, the product information corresponding to the store is stored.

"Category information" indicates information about the category of goods traded in the store. "Category information" includes information indicating the category of goods traded in the store. The "category information" includes information indicating a main category, which is a category of products that the store mainly trades (mainly), a subcategory that is a category of products that the store trades other than the main category, and the like.

For example, the main category and the sub-category may be specified by the store operator, or may be determined from the sales of the store. Further, there may be a plurality of main categories, and there may be no sub-categories. The above is only an example, and the "category information" may include various information about the category. In the example of FIG. 9, although it is indicated by an abstract code such as “category information # 1”, the category information corresponding to the store is stored.

"Sales information" indicates information about the sales of the store. "Sales information" includes information indicating sales in the store. The "sales information" includes information indicating the number of products sold in the store and information indicating the sales amount of the products in the store. For example, "sales information" includes information indicating sales for each period such as one day, one week, one month, and the like. The "sales information" may include various information related to sales, such as information indicating sales for each product traded in the store and information indicating sales for each category.

The "sales information" may include information such as the number of sales and the sales amount for each product. "Sales information" may include various information related to the number of sales and the amount of sales, such as the number of products sold and the value of the sales amount, the ratio of the number of sales to the number of products handled, and the aggregation period of the number and amount. good. The above is only an example, and the "sales information" may include various information related to sales. In the example of FIG. 9, although it is indicated by an abstract code such as “sales information # 1”, the sales information corresponding to the store is stored.

"Evaluation information" indicates information on evaluation of a business operator, a store opened by the business operator, and products traded in the store. The "evaluation information" includes information indicating the evaluation from the user. For example, "evaluation information" includes the number of daily visits (page views) to the page corresponding to the store of the commercial transaction site, the number of store reviews indicating the number of reviews posted by the user to the store, and the number of posts by the user. Information such as the store score that evaluated the store in the review was included. The number of page views is not limited to one day, and may include the number of page views for a predetermined period (one week, etc.). For example, the number of store reviews may be the average number of store reviews per day in the store.

For example, the store score may be in five stages from the lowest score "1" to the highest score "5". The store score may be the average score of the reviews posted by the users before the store closes. As for the number of store reviews and the store score, the total number and the average of the scores in a predetermined period (for example, one week) may be stored, the cumulative number up to the present time may be stored, or the cumulative number up to the present time may be stored. Information aggregated in various ways, such as the cumulative number divided by the number of years the store lasts, may be stored. For example, the "evaluation information" may include information indicating the ratio of the number of reviews posted to the page view.

For example, the "evaluation information" includes information such as the number of product reviews indicating the number of reviews posted on the product, the product score indicating the score in the reviews posted by the user on the product, and the like. For example, the product score may be in five stages from the lowest score "1" to the highest score "5". In addition, for example, the product score may be the average score of the scores in the reviews posted by the users by the end of the handling of the products in the store. The above is only an example, and the "evaluation information" may include various information related to the evaluation. In the example of FIG. 9, although it is indicated by an abstract code such as “evaluation information # 1”, the evaluation information corresponding to the store is stored.

"Score information" indicates score information for the store. The "score information" includes various information estimated by the information processing apparatus 100. For example, the score output by the model is stored in the "score information". As shown in FIG. 9, for stores that are not subject to score calculation, such as stores that have already closed, the score information item is blank (“−” (hyphen)).

"Closed" indicates the probability (possibility) that the store will leave the commercial transaction site. In the "closed store", a store exit score indicating the probability (possibility) of the store exiting is stored. For example, in "closed store", a score (closed store score) output by the closed store model (for example, model M1) is stored. In the example of FIG. 9, although it is indicated by an abstract code such as “store exit score # 1”, a specific value (numerical value) such as “0.7” is stored.

"Future proof" indicates the future of the store. "Future Proof" stores a future score that indicates the future of the store. For example, the “future potential” stores a score (future potential score) output by a future model (for example, model M2). In the example of FIG. 9, although it is indicated by an abstract code such as “future potential score # 1”, specific values (numerical values) such as “0.5” and “10” are stored.

The store store information storage unit 121 is not limited to the above, and may store various information depending on the purpose. For example, the store store information storage unit 121 may store information indicating the reason for closing the store that has closed. For example, the reason for leaving the store is used to distinguish correct answer information in the model generation process described later. For example, the store owner information storage unit 121 leaves information such as forced closure indicating that the store has been forcibly closed from the commercial transaction site and closure indicating that the store has been voluntarily closed by the business operator. It may be stored as information indicating the reason for the store.

(Model information storage unit 122)
The model information storage unit 122 according to the embodiment stores information about the model. For example, the model information storage unit 122 stores information (model data) of a learned model (model) learned (generated) by the learning process. The model information storage unit 122 stores the data (learning data) used for learning in association with the trained model (model). FIG. 10 is a diagram showing an example of a model information storage unit according to an embodiment. In the example shown in FIG. 10, the model information storage unit 122 includes items such as "model ID", "use", "model data", and "learning data". In the example of FIG. 10, the model information storage unit 122 stores the data (learning data) used for learning in association with the trained model (model).

The "model ID" indicates identification information for identifying the model. "Use" indicates the use of the corresponding model. "Model data" indicates model data. In FIG. 10 and the like, an example in which conceptual information such as "MDT1" is stored in "model data" is shown, but in reality, the model is configured such as information on model configuration (network configuration) and information on parameters. Contains various information. For example, "model data" includes information including nodes at each layer of the network, functions adopted by each node, connection relationships of the nodes, and connection coefficients set for connections between the nodes.

"Training data" indicates data used for training a trained model (model). In the "training data", information indicating the data set used for training the corresponding model is stored. For example, the "learning data" is stored as learning data (also referred to as "learning data") by associating the data (input information) with the correct answer information (output information) corresponding to the data. In FIG. 10, an example in which conceptual information such as "LDT1" is stored in "learning data" is shown, but in reality, data (input information) and correct answer information (output information) corresponding to the data are supported. Contains various information about the data used to train the model.

In FIG. 10, the model (model M1) identified by the model ID “M1” indicates that the use is “closed”. That is, it is shown that the model M1 is a model that outputs information (score) indicating the probability (possibility) that the store corresponding to the input data leaves the commercial transaction site. Further, it is shown that the model data of the model M1 is the model data MDT1. Further, it is shown that the training data used for training the model M1 is the training data LDT1.

Further, the model (model M2) identified by the model ID "M2" indicates that the use is "future potential". That is, it is shown that the model M2 is a model that outputs information (score) indicating the future potential of the store corresponding to the input data. For example, the model M2 outputs information (score) for estimating a change in store sales corresponding to the input data. The model data of the model M2 indicates that it is the model data MDT2. Further, it is shown that the training data used for training the model M2 is the training data LDT2.

The model (model M11) identified by the model ID "M11" indicates that the use is "January retreat store". That is, the model M11 indicates that the store corresponding to the input data is a model that outputs information (score) indicating the probability (possibility) of leaving the store within one month from the commercial transaction site. Further, it is shown that the model data of the model M11 is the model data MDT11. Further, it is shown that the training data used for training the model M11 is the training data LDT11.

The model (model M12) identified by the model ID "M12" indicates that the use is "February retreat store". That is, the model M12 indicates that the store corresponding to the input data is a model that outputs information (score) indicating the probability (possibility) of leaving the store within two months from the commercial transaction site. Further, it is shown that the model data of the model M12 is the model data MDT12. Further, it is shown that the training data used for training the model M12 is the training data LDT12.

The model (model M13) identified by the model ID "M13" indicates that the use is "March retreat store". That is, the model M13 indicates that the store corresponding to the input data is a model that outputs information (score) indicating the probability (possibility) of leaving the store within three months from the commercial transaction site. Further, it is shown that the model data of the model M13 is the model data MDT13. Further, it is shown that the training data used for training the model M13 is the training data LDT13.

Although up to model M13 is shown in FIG. 10, the model information storage unit 122 estimates (predicts) a model that estimates (predicts) the probability of store closure corresponding to various periods, such as model M14 that predicts a store retreat in April. You may remember it.

The model (model M31) identified by the model ID "M31" indicates that the use is "number of stores closed". That is, it is shown that the model M31 is a model that outputs information (score) indicating the number of stores (number of stores closed) that are closed from the commercial transaction site. Further, it is shown that the model data of the model M31 is the model data MDT31. Further, it is shown that the training data used for training the model M31 is the training data LDT31.

The model M31 is only an example, and the model information storage unit 122 may store not only the model M31 used for estimating the number of store closures but also a model for estimating (predicting) the number of occurrences of various negative actions. ..

The model (model M41) identified by the model ID "M41" indicates that the use is "category sales". That is, it is shown that the model M41 is a model that outputs information (score) indicating sales on a commercial transaction site for products in the category corresponding to the input data. Further, it is shown that the model data of the model M41 is the model data MDT41. Further, it is shown that the training data used for training the model M41 is the training data LDT41. When the model is learned for each category, the model information storage unit 122 may store the model for each category in association with the category.

The model information storage unit 122 is not limited to the above, and may store various information depending on the purpose.

(Threshold information storage unit 123)
The threshold information storage unit 123 according to the embodiment stores various information regarding the threshold value. FIG. 11 is a diagram showing an example of the threshold information storage unit according to the embodiment. The threshold information storage unit 123 shown in FIG. 11 includes items such as “threshold value”, “value”, and “use”.

The "threshold value" indicates information for identifying the threshold value. The "value" indicates a specific numerical value (value) of the corresponding threshold value. "Use" refers to a use for which the corresponding threshold is used.

In the example of FIG. 11, the value of the threshold value (threshold value TH1) identified by the threshold value “TH1” is shown to be “0.9”. The use of the threshold TH1 indicates that it is a grouping (classification). That is, the threshold value TH1 indicates that it is the first threshold value used for grouping (classification) of stores. For example, the threshold TH1 is the first threshold used to classify the stores of the first group.

The value of the threshold value (threshold value TH2) identified by the threshold value “TH2” indicates that it is “0.8”. The use of the threshold TH2 indicates that it is a grouping (classification). That is, the threshold value TH2 indicates that it is the second threshold value used for grouping (classification) of stores. For example, the threshold TH2 is a second threshold used to classify stores in the second group.

The value of the threshold value (threshold value TH11) identified by the threshold value “TH11” indicates that it is “VL11”. It is shown that the use of the threshold value TH11 is to estimate the closing time. That is, the threshold value TH11 indicates that the threshold value is used to estimate when the store will leave the commercial transaction site. In the example of FIG. 11, although it is indicated by an abstract reference numeral such as “VL11”, the value of each threshold value is a specific numerical value according to the intended use.

The value of the threshold value (threshold value TH21) identified by the threshold value “TH21” indicates that it is “VL21”. It is shown that the use of the threshold value TH21 is to determine the proposal destination. That is, the threshold value TH21 indicates that the threshold value is used to determine the store in which the proposal is made.

The threshold information storage unit 123 is not limited to the above, and may store various information depending on the purpose.

(Group information storage unit 124)
The group information storage unit 124 according to the embodiment stores various information about the group. The group information storage unit 124 stores various information about a plurality of groups corresponding to each of the plurality of groups. FIG. 12 is a diagram showing an example of a group information storage unit according to an embodiment. The group information storage unit 124 shown in FIG. 12 includes items such as "group ID", "group information", and "member".

The "group ID" indicates identification information for identifying the group. "Group information" indicates group information in the corresponding group. For example, "group information" indicates group information including information on index values (number of months) set in the group. In FIG. 12, an example in which conceptual information such as “SM1” is stored in “group information” is shown, but in reality, specific information such as an index value (months) set in the group is stored. To.

"Member" refers to a member of the group. For example, information (store ID, etc.) that identifies a member (store) belonging to a group is stored in the "member".

In FIG. 12, it is shown that the group information of the group (first group) identified by the group ID “GP1” is the group information SM1. In FIG. 12, it is shown that “3” is set as an index value in the group GP1 which is the first group. For example, the number of months "3" is set as an index value in the group GP1 which is the first group. It is shown that stores SP2, SP12, etc. belong to the group GP1 as members.

The group information storage unit 124 is not limited to the above, and may store various information depending on the purpose.

(Social Situation Information Memory Department 125)
The social situation information storage unit 125 according to the embodiment stores various information related to the social situation. In the example of FIG. 13, social situation information regarding the economy is stored. FIG. 13 is a diagram showing an example of the social situation information storage unit according to the embodiment. The social situation information storage unit 125 shown in FIG. 13 includes items such as "social situation ID", "time point", and "social situation information".

The "social situation ID" indicates identification information for identifying the social situation (object). In addition, character information or the like for specifying the social situation corresponding to the social situation ID may be stored in association with each social situation ID. "Time point" indicates the time (time point) in which the corresponding social situation was. For example, "time point" indicates the time (time point) when the corresponding social situation information was acquired. In the example of FIG. 8, the time point is indicated by an abstract code such as "time point # 1", but information indicating the corresponding period such as "19XY year Z month ..." in parentheses is included. May be good.

"Social situation information" indicates information about the social situation identified by the social situation ID. In the example of FIG. 13, the social situation information is indicated by an abstract code such as "SDT1", but each social situation information relates to the economy (economic situation) at the time corresponding to the social situation identified by the social situation ID. Information indicating various social conditions such as information is included. For example, social situation information includes various information indicating economic conditions such as stock prices, unemployment rate, and interest rates.

In the example of FIG. 13, the social situation identified by the social situation ID “SC1” indicates that the social situation corresponds to the time point # 1. The social situation information of the social situation identified by the social situation ID "SC1" indicates that it is the social situation information SDT1.

The social situation information storage unit 125 is not limited to the above, and may store various information depending on the purpose.

(Commercial transaction site information storage unit 126)
The commercial transaction site information storage unit 126 according to the embodiment stores various information related to the commercial transaction site. For example, the commercial transaction site information storage unit 126 stores statistical information about the commercial transaction site, such as total sales on the commercial transaction site. FIG. 14 is a diagram showing an example of a commercial transaction site information storage unit according to an embodiment. The commercial transaction site information storage unit 126 shown in FIG. 14 includes items such as “category” and “sales information by category”. In the example of FIG. 14, the case where the sales information for each category is stored is shown as an example, but the commercial transaction site information storage unit 126 may store various information, not limited to the sales information for each category.

"Category" indicates identification information for identifying a category. In FIG. 14, an example in which conceptual information such as "category # 1" is stored in "category" is shown, but in reality, for specifying a category such as a category name or a category identifier (category ID). Information is stored.

"Sales information by category" indicates the sales of products belonging to the corresponding category on the commercial transaction site. In FIG. 14, an example in which conceptual information such as “category sales # 1” is stored in “sales information by category” is shown, but in reality, information indicating sales in the corresponding category is stored. The sales information by category is not limited to the sales of the category, and may include various information related to the sales in the category such as the number of sales of the category.

In FIG. 14, it is shown that the sales of the category # 1 are the category sales # 1. For example, the sales of the goods belonging to the category # 1 on the commercial transaction site are the values indicated by the category sales # 1. Further, the sales of the category # 2 are the category sales # 2, and the sales of the category # 3 are the category sales # 3.

The commercial transaction site information storage unit 126 is not limited to the above, and may store various information depending on the purpose.

(Factoring information storage unit 127)
The factoring information storage unit 127 according to the embodiment stores various information related to factoring. FIG. 15 is a diagram showing an example of a factoring information storage unit according to an embodiment. The factoring information storage unit 127 shown in FIG. 15 includes items such as “store” and “factoring information”.

“Store” indicates identification information for identifying the store. "Factoring information" indicates information (factoring information) regarding factoring of the corresponding store. For example, "factoring information" indicates whether or not the corresponding store factoring is in use. For example, a store whose "factoring information" is "-" (hyphen) indicates that the store does not use factoring.

A store in which "factoring information" is factoring information # * (* is an arbitrary number) indicates that the store is using factoring. FIG. 15 shows an example in which conceptual information such as "factoring information # 12" is stored in "factoring information", but in reality, the amount of factoring and its collection mode (collection period, collection interval, etc.), etc. Contains specific information about factoring.

FIG. 15 shows that store SP1 does not utilize factoring. Further, the store SP12 indicates that factoring is being used and the factoring information is factoring information # 12.

The factoring information storage unit 127 is not limited to the above, and may store various information depending on the purpose.

(Control unit 130)
Returning to the description of FIG. 8, the control unit 130 is a controller, and is stored in a storage device inside the information processing device 100 by, for example, a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). Various programs (corresponding to an example of an information processing program) are realized by executing the RAM as a work area. Further, the control unit 130 is a controller, and is realized by, for example, an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 8, the control unit 130 includes an acquisition unit 131, a learning unit 132, an estimation unit 133, a classification unit 134, a determination unit 135, a determination unit 136, and a provision unit 137. Realize or execute the functions and actions of information processing described below. The internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 8, and may be any other configuration as long as it is configured to perform information processing described later.

(Acquisition unit 131)
The acquisition unit 131 acquires various information. The acquisition unit 131 acquires various information from an external device such as the user terminal 10. Further, the acquisition unit 131 acquires various information from the storage unit 120. Further, the acquisition unit 131 includes a store store information storage unit 121, a model information storage unit 122, a threshold information storage unit 123, a group information storage unit 124, a social situation information storage unit 125, and a commercial transaction site information storage unit 126. And, various information is acquired from the factoring information storage unit 127 and the like.

The acquisition unit 131 receives various information from an external information processing device via the communication unit 110. The acquisition unit 131 stores the received information in the storage unit 120. The acquisition unit 131 receives various information from the user terminal 10, the store device 20, or the web server 30. The acquisition unit 131 acquires information from the user terminal 10 used by the user. The acquisition unit 131 acquires various information from the store device 20. For example, the acquisition unit 131 acquires information about a store opened on the commercial transaction site by the store owner from the store device 20 used by the store owner. For example, the acquisition unit 131 acquires various information from the web server 30. For example, the acquisition unit 131 acquires information about the commercial transaction site from the web server 30. For example, the acquisition unit 131 acquires information about the store of the commercial transaction site from the web server 30.

The acquisition unit 131 acquires store information about each of the plurality of stores that open on the commercial transaction site on the network. The acquisition unit 131 acquires a learning model (model). The acquisition unit 131 acquires store sales information indicating the past sales of the target store, which is a store that opens on a commercial transaction site on the network. The acquisition unit 131 acquires category sales information indicating past sales for each category of products handled on the commercial transaction site. The acquisition unit 131 acquires store sales information indicating changes in past sales of the target store, which is a store that opens on a commercial transaction site on the network. The acquisition unit 131 acquires category sales information showing changes in past sales for each category of products handled on the commercial transaction site.

The acquisition unit 131 acquires a first learning model that predicts the probability of occurrence of a negative event, which is an event caused by a store opened on a commercial transaction site on the network and is a negative event for the commercial transaction site. The acquisition unit 131 acquires a second learning model that predicts the future possibility of the sales of the store opened on the commercial transaction site. The acquisition unit 131 acquires the first learning model that outputs the score indicating the probability that a negative event occurs due to the store by inputting the store information of the store.

The acquisition unit 131 acquires the second learning model that outputs the score indicating the future possibility of the sales of the store by inputting the store information of the store. The acquisition unit 131 acquires a first learning model that predicts the probability that the store's receivables will become uncollectible. The acquisition unit 131 acquires a first learning model that predicts the probability of store closure from a commercial transaction site.

The acquisition unit 131 acquires store information about each of the plurality of stores that open on the commercial transaction site on the network. The acquisition unit 131 acquires store information including information indicating whether or not each of the plurality of stores is advertising. The acquisition unit 131 acquires store information including information indicating the products handled, which are the products handled by each of the plurality of stores.

(Learning Department 132)
The learning unit 132 executes a learning process for learning a learning model (model). For example, the learning unit 132 executes the learning process based on various information acquired by the acquisition unit 131. The learning unit 132 executes the learning process based on the information from the external information processing device and the information stored in the storage unit 120. The learning unit 132 executes the learning process based on the information stored in the model information storage unit 122. The learning unit 132 stores the model generated by learning in the model information storage unit 122.

The learning unit 132 learns a learning model for predicting the probability of occurrence of a negative event, which is an event that occurs within a predetermined period by a store opened on a commercial transaction site on the network and is a negative event for the commercial transaction site. The learning unit 132 learns a learning model that inputs a store information of the store and outputs a score indicating the probability that a negative event occurs due to the store.

The learning unit 132 learns a learning model that predicts the probability that a store's receivables will become uncollectible. The learning unit 132 learns a learning model that predicts the probability of store exits from a commercial transaction site.

The learning unit 132 learns a plurality of learning models, each of which corresponds to a different period, and which predicts the probability that a negative event will occur due to the store within the corresponding period. The learning unit 132 learns a plurality of learning models that output scores indicating the probability that a negative event will occur due to the store within the corresponding period.

The learning unit 132 performs a learning process. The learning unit 132 performs various learning. The learning unit 132 learns various types of information based on the information acquired by the acquisition unit 131. The learning unit 132 learns (generates) a model. The learning unit 132 learns various information such as a model. The learning unit 132 generates a model by learning. The learning unit 132 learns a model by using various techniques related to machine learning. For example, the learning unit 132 learns the parameters of the model (network). The learning unit 132 learns a model by using various techniques related to machine learning.

The learning unit 132 generates various learning models such as models M1, M2, M11, M12, M13, M21, M22, M31, and M41. The learning unit 132 learns the parameters of the network. For example, the learning unit 132 learns the network parameters of various learning models such as models M1, M2, M11, M12, M13, M21, M22, M31, and M41. The learning unit 132 performs learning processing using the learning data stored in the model information storage unit 122 to perform various learning processes such as models M1, M2, M11, M12, M13, M21, M22, M31, and M41. Generate a learning model. For example, the learning unit 132 generates a model used for speech recognition. The learning unit 132 learns the network parameters of various learning models such as models M1, M2, M11, M12, M13, M21, M22, M31, and M41, thereby learning the models M1, M2, M11, M12, and M13. Various learning models such as M21, M22, M31, and M41 are generated.

The learning unit 132 performs learning processing based on the learning data (teacher data) stored in the model information storage unit 122. The learning unit 132 performs learning processing using the learning data stored in the model information storage unit 122 to perform various learning processes such as models M1, M2, M11, M12, M13, M21, M22, M31, and M41. Generate a learning model.

For example, the learning unit 132 generates a model used for estimating the probability (possibility) of store closure. For example, the learning unit 132 generates a model that outputs a score indicating the probability (possibility) that the store corresponding to the store information leaves the commercial transaction site when the store information of the store is input. The learning unit 132 generates the model M1 by learning the parameters of the network of the model M1.

The learning method by the learning unit 132 is not particularly limited, but for example, learning data in which data (input information) and its correct answer information (output information) are linked is prepared, and the learning data is based on a multi-layer neural network. You may learn by inputting to the calculation model. Further, for example, a method based on DNN (Deep Neural Network) such as CNN (Convolutional Neural Network) or 3D-CNN may be used. When targeting time-series data such as voice, the learning unit 132 uses a method based on a recurrent neural network (RNN) or LSTM (Long Short-Term Memory units) that extends RNN. May be good.

(Estimation unit 133)
The estimation unit 133 executes an estimation process for estimating various information. The estimation unit 133 stores the information estimated by the estimation process in the storage unit 120. For example, the estimation unit 133 executes the estimation process based on various information acquired by the acquisition unit 131. The estimation unit 133 executes the estimation process based on various information stored in the storage unit 120. The estimation unit 133 executes estimation processing based on various information received from an external information processing device.

The estimation unit 133 executes the estimation process using the model learned by the learning unit 132. For example, the estimation unit 133 executes the estimation process based on the information classified by the classification unit 134. For example, the estimation unit 133 executes the estimation process based on the information determined by the determination unit 135. For example, the estimation unit 133 executes the estimation process based on the information determined by the determination unit 136.

The estimation unit 133 estimates the number of occurrences of negative events, which are events caused by stores opened on commercial trading sites on the network and which are negative events for commercial trading sites, using social situation information indicating the economic situation of society. .. The estimation unit 133 estimates the number of uncollectible receivables in the store as the number of occurrences. The estimation unit 133 estimates the number of stores that leave the commercial transaction site as the number of occurrences. The estimation unit 133 estimates the number of stores that leave the commercial transaction site by using a learning model that learns the characteristics of the relationship between the economic situation of the society and the number of stores that leave the commercial transaction site.

The estimation unit 133 estimates the closing probability, which is the probability that the target store, which is a store opening on the commercial transaction site on the network, will leave the commercial transaction site after a predetermined period. The estimation unit 133 estimates the store exit probability of the target store by using a learning model that learns the characteristics of the relationship between the store opening on the commercial transaction site and the exit from the commercial transaction site.

The estimation unit 133 uses a plurality of learning models, each of which corresponds to a different period, and a plurality of learning models for estimating the probability that the store will close within the corresponding period, and the store exit probability of the target store. To estimate. The estimation unit 133 estimates the store exit probability of the target store by using a plurality of learning models that output scores indicating the probability that the store will be closed within the corresponding period. The estimation unit 133 updates the closing probability of the target store when the collection of the accounts receivable corresponding to the future receivables of the target store satisfies the condition regarding the delay.

(Classification unit 134)
The classification unit 134 executes a classification process for classifying various information. The classification unit 134 stores the information classified by the classification process in the storage unit 120. For example, the classification unit 134 executes the classification process based on various information acquired by the acquisition unit 131. The classification unit 134 executes the classification process based on various information stored in the storage unit 120. The classification unit 134 executes the classification process based on various information received from the external information processing device.

The classification unit 134 executes the classification process using the model learned by the learning unit 132. For example, the classification unit 134 executes the classification process based on the information determined by the determination unit 136. For example, the classification unit 134 executes the classification process based on the information determined by the determination unit 135. For example, the classification unit 134 executes the classification process based on the information estimated by the estimation unit 133.

The classification unit 134 classifies the plurality of stores into a plurality of groups based on the store information regarding each of the plurality of stores opened on the commercial transaction site on the network. The classification unit 134 classifies the plurality of stores into a plurality of groups according to the risk of uncollectible receivables of the plurality of stores based on the store information. The classification unit 134 classifies a plurality of stores into a plurality of groups according to the possibility of leaving the store from the commercial transaction site.

The classification unit 134 classifies a plurality of stores into a plurality of groups by using a learning model that learns the characteristics of the relationship between the stores that open on the commercial transaction site and the exits from the commercial transaction site. The classification unit 134 classifies a plurality of stores into a plurality of groups based on store information including information on products sold by each of the plurality of stores. The classification unit 134 classifies a plurality of stores into a plurality of groups based on the store information including the sales information of each of the plurality of stores. The classification unit 134 classifies a plurality of stores opening on the commercial transaction site into a plurality of groups by using a threshold value based on the number of occurrences estimated by the estimation unit 133.

(Judgment unit 135)
The determination unit 135 executes a determination process for determining various types of information. For example, the determination unit 135 executes the determination process based on various information acquired by the acquisition unit 131. For example, the determination unit 135 executes the determination process based on the information stored in the storage unit 120.

The determination unit 135 executes the determination process using the model learned by the learning unit 132. For example, the determination unit 135 executes the determination process based on the information estimated by the estimation unit 133. For example, the determination unit 135 executes the determination process based on the information classified by the classification unit 134. For example, the determination unit 135 executes the determination process based on the information determined by the determination unit 136.

The determination unit 135 determines the tendency of sales for each category based on the past sales for each category. The determination unit 135 determines the tendency of sales for each category based on the transition of past sales for each category. For example, the determination unit 135 determines the sales tendency for each category using the score output by the model M41. For example, when the score output by the model M41 is equal to or higher than a predetermined threshold value (for example, 1.2 or the like), the determination unit 135 shows an upward trend in the sales tendency of the category corresponding to the model input information input to the model M41. Judge that there is. For example, when the score output by the model M41 is less than a predetermined threshold value (for example, 0.8, etc.), the determination unit 135 shows a downward trend in the sales tendency of the category corresponding to the model input information input to the model M41. Judge that there is.

The determination unit 135 may execute a generation process for generating various information. The determination unit 135 stores the generated information in the storage unit 120. For example, the determination unit 135 executes the generation process based on various information acquired by the acquisition unit 131. The determination unit 135 executes the generation process based on various information stored in the storage unit 120. The determination unit 135 executes the generation process based on various information received from the external information processing device.

For example, the determination unit 135 executes the generation process based on the information determined by the determination unit 136. For example, the determination unit 135 executes the generation process based on the information determined by the determination unit 135. For example, the determination unit 135 executes the generation process based on the information classified by the classification unit 134. For example, the determination unit 135 executes the generation process based on the information estimated by the estimation unit 133.

The determination unit 135 generates the content. The determination unit 135, for example, uses various techniques such as Java (registered trademark) as appropriate to generate a screen (content) to be provided to the user terminal 10. The determination unit 135 may generate a screen (content) to be provided to the user terminal 10 based on the format of CSS, Javascript (registered trademark), or HTML. Further, for example, the determination unit 135 may generate a screen (content) in various formats such as JPEG (Joint Photographic Experts Group), GIF (Graphics Interchange Format), and PNG (Portable Network Graphics).

(Decision unit 136)
The determination unit 136 executes a determination process for determining various information. The determination unit 136 stores the determined information in the storage unit 120. For example, the determination unit 136 executes the determination process based on various information acquired by the acquisition unit 131. The determination unit 136 executes the determination process based on various information stored in the storage unit 120. For example, the determination unit 136 executes the determination process based on various information received from the external information processing apparatus.

The determination unit 136 executes the determination process using the model learned by the learning unit 132. For example, the determination unit 136 executes the determination process based on the information estimated by the estimation unit 133. For example, the determination unit 136 executes the determination process based on the information classified by the classification unit 134. For example, the determination unit 136 executes the determination process based on the information determined by the determination unit 135.

For example, the determination unit 136 calculates various information based on the information estimated by the estimation unit 133. For example, the determination unit 136 calculates various information based on the information determined by the determination unit 135. For example, the determination unit 136 calculates various information based on the information generated by the determination unit 135. For example, the determination unit 136 calculates various information based on the information classified by the classification unit 134. The determination unit 136 calculates various scores using the model. For example, the determination unit 136 calculates various scores.

The determination unit 136 determines the conditions for the purchase of receivables for each group classified by the classification unit 134. The decision unit 136 determines the conditions that are more favorable to the stores belonging to the group as the risk is lower. The determination unit 136 determines conditions that are more favorable to the store belonging to the group, as the group is less likely to leave the store.

The decision-making unit 136 determines the conditions for the purchase of future receivables for each group. The determination unit 136 determines the index value for the purchase of future receivables for each group. The determination unit 136 determines the upper limit of the purchase of future receivables of one store based on the index value of the group to which one store belongs and the sales of one store. The determination unit 136 determines the purchase upper limit of one store based on the value obtained by multiplying the index value of the group to which one store belongs and the sales of one store. The determination unit 136 determines the number of months for which future claims are to be purchased as an index value for each group.

The determination unit 136 determines the conditions for purchasing the receivables of the target store based on the probability of occurrence of a negative event by the target store predicted by using the learning model learned by the learning unit 132. The determination unit 136 determines the conditions that are advantageous to the target store as the probability that the receivables of the target store become uncollectible is low. The determination unit 136 determines the conditions that are more advantageous to the target store as the probability of the target store closing occurs is lower.

The determination unit 136 determines the conditions for the purchase of future receivables of the target store based on the probability of occurrence of a negative event by the target store. The determination unit 136 determines the index value for the purchase of future receivables of the target store. The determination unit 136 determines the upper limit of the purchase of future receivables of the target store based on the index value of the target store and the sales of the target store. The determination unit 136 determines the purchase upper limit of the target store based on the value obtained by multiplying the index value of the target store and the sales of the target store. The determination unit 136 determines the number of months for which future receivables of the target store are to be purchased as an index value.

The determination unit 136 uses a plurality of learning models to determine the number of months for which future receivables of the target store are to be purchased as an index value. The determination unit 136 determines the number of months for the purchase of future receivables of the target store based on the period corresponding to the learning model having the shortest period among the learning models whose scores are equal to or higher than the threshold value.

The determination unit 136 determines the conditions for the purchase of the receivables based on the number of occurrences estimated by the estimation unit 133.

The determination unit 136 determines the receivables of the target store based on the sales tendency of the target store indicated by the store sales information and the sales tendency of the category corresponding to the handling product which is the product handled by the target store indicated by the category sales information. Determine the maximum purchase amount. The determination unit 136 is based on the trend of the sales transition of the target store indicated by the store sales information and the trend of the sales transition of the category corresponding to the product handled by the target store indicated by the category sales information. Determine the maximum purchase amount of receivables. The determination unit 136 determines the target category corresponding to the target store based on the products handled by the target store, and determines the maximum purchase amount of the receivables of the target store based on the trend of the sales transition of the determined target category.

The decision unit 136 determines a plurality of target categories corresponding to the target store based on the products handled by the target store, and purchases the receivables of the target store based on the tendency of the sales transition of each of the determined target categories. Determine the maximum amount. The determination unit 136 determines the maximum purchase amount of the receivables of the target store based on the weighting according to the number of products of the plurality of target categories on the commercial transaction site. The determination unit 136 determines the maximum purchase amount of the receivables of the target store by increasing the weighting as the category has a large number of products handled on the commercial transaction site.

The determination unit 136 determines the main category among the plurality of target categories, and determines the maximum purchase amount of the receivables of the target store based on the trend of the sales transition of the determined main category. The determination unit 136 determines the maximum purchase amount of the receivables of the target store by increasing the weight of the main category more than the category different from the main category among the plurality of target categories.

The decision unit 136 learns the characteristics of the relationship between the trend of the store sales transition up to a predetermined time point, the trend of the sales transition of the category corresponding to the product handled by the store, and the sales of the store after the predetermined time point. Using the learning model that was created, the maximum purchase amount of the receivables of the target store is determined.

The determination unit 136 determines the maximum purchase amount of future receivables of the target store. The determination unit 136 determines the index value for the purchase of future receivables of the target store based on the trend of the sales transition of the target store and the trend of the sales transition of the category corresponding to the products handled by the target store, and the determined index. The value is used to determine the maximum purchase limit for future receivables of the target store. The determination unit 136 determines the maximum purchase amount of the receivables of the target store according to the determination result by the determination unit 135.

For example, when the determination unit 135 determines that the sales in the category of the target store are on an upward trend, the determination unit 136 may increase the purchase upper limit of the receivables of the target store. In this case, the determination unit 136 may increase the purchase upper limit of the receivables of the target store when the determination unit 135 determines that the sales of the main category of the target store are on an upward trend. Further, for example, when the determination unit 135 determines that the sales in the category of the target store are in a downward trend, the determination unit 136 may reduce the upper limit of the purchase of the receivables of the target store. In this case, the determination unit 136 may reduce the purchase upper limit of the receivables of the target store when the determination unit 135 determines that the sales of the main category of the target store are on a downward trend.

The decision unit 136 determines the receivables of the target store based on the probability of occurrence of a negative event by the target store predicted using the first learning model and the future possibility of sales of the target store predicted using the second learning model. Determine the conditions for purchasing. The determination unit 136 determines the conditions that are advantageous to the target store as the probability that the receivables of the target store become uncollectible is low. The determination unit 136 determines the conditions that are more advantageous to the target store as the probability of the target store closing occurs is lower.

The determination unit 136 determines the conditions for purchasing the receivables of the target store by using the first learning model in which the characteristics of the relationship between the store opened on the commercial transaction site and the exit from the commercial transaction site are learned. The determination unit 136 determines the conditions for purchasing future receivables of the target store based on the probability of occurrence of a negative event by the target store and the future possibility of sales of the target store.

The determination unit 136 determines the mode of collecting accounts receivable corresponding to the future receivables of the target store based on the closing probability of the target store estimated by the estimation unit 133. The determination unit 136 determines the collection rate of accounts receivable corresponding to the future receivables of the target store based on the closing probability of the target store. The determination unit 136 determines the mode of collecting accounts receivable corresponding to the future receivables of the target store based on the closing probability of the target store for each period.

The determination unit 136 determines the mode of collecting accounts receivable corresponding to the future receivables of the target store based on the period corresponding to the learning model having the shortest period among the learning models whose scores are equal to or higher than the threshold value. The determination unit 136 determines the period corresponding to the learning model with the shortest period among the learning models whose score is equal to or higher than the threshold value as the collection period of accounts receivable corresponding to the future receivables of the target store, and according to the collection period. , Determine the collection rate of accounts receivable corresponding to future receivables of the target store.

If the collection of accounts receivable corresponding to the future receivables of the target store satisfies a predetermined condition, the determination unit 136 updates the collection mode of the accounts receivable corresponding to the future receivables of the target store. If the collection of accounts receivable corresponding to the future receivables of the target store satisfies the condition regarding delay, the determination unit 136 updates the collection mode of the accounts receivable corresponding to the future receivables of the target store. The determination unit 136 updates the collection mode of the accounts receivable corresponding to the future receivables of the target store based on the updated probability of closing the target store.

The determination unit 136 determines, among the plurality of stores, the proposed store to which the purchase of the receivable is proposed, based on the store information regarding each of the plurality of stores. The decision unit 136 determines the proposal destination store depending on whether or not the advertisement is advertised. The decision unit 136 determines the store that does not advertise as the proposed store.

The decision unit 136 determines the proposal destination store according to the products handled. The determination unit 136 determines a store in which the number of products handled is less than a predetermined threshold value as the proposal destination store.

The decision unit 136 decides the proposed store using a learning model that learns the characteristics of the relationship between the store that opens on the commercial transaction site and the change in sales when the store takes measures to increase sales. do. The determination unit 136 determines the proposal destination store by using the learning model that outputs the score indicating the change in sales when the measure is taken by inputting the store information of the store. The determination unit 136 determines a store whose score output by the learning model is equal to or higher than a predetermined threshold value as the proposal destination store.

(Providing Department 137)
The providing unit 137 provides various information. The providing unit 137 transmits various information to an external information processing device via the communication unit 110. The providing unit 137 transmits various information to the user terminal 10, the store device 20, and the web server 30. The providing unit 137 transmits the content to the user terminal 10.

The providing unit 137 provides the information determined by the determining unit 136. The providing unit 137 provides the information determined by the determination unit 135. The providing unit 137 provides the information generated by the determination unit 135. The providing unit 137 provides the information classified by the classification unit 134. The providing unit 137 provides the information estimated by the estimation unit 133.

The providing unit 137 transmits the information determined by the determining unit 136 to the store device 20 used by the target store. The providing unit 137 provides information indicating the conditions determined by the determining unit 136. The providing unit 137 provides information indicating the purchase upper limit amount determined by the determining unit 136. The providing unit 137 provides information indicating the collection mode of the accounts receivable corresponding to the future receivables of the target store determined by the determining unit 136. The providing unit 137 provides information for proposing the purchase of receivables to the proposed store determined by the determining unit 136.

[3. Processing procedure]
Next, the procedure of information processing executed by the information processing apparatus 100 according to the embodiment will be described with reference to FIGS. 16 to 22. 16 to 22 are flowcharts showing an example of the flow of information processing executed by the information processing apparatus.

First, FIG. 16 will be described. For example, FIG. 16 shows an example of determination of conditions regarding the purchase of receivables for each group performed by the information processing apparatus 100. In FIG. 16, the information processing apparatus 100 classifies the plurality of stores into a plurality of groups based on the store information regarding each of the plurality of stores opened on the commercial transaction site on the network (step S101). Then, the information processing apparatus 100 determines the conditions for purchasing the receivables for each group (step S102).

Next, FIG. 17 will be described. For example, FIG. 17 shows an example of learning a model performed by the information processing apparatus 100 and processing using the learned model. In FIG. 17, the information processing apparatus 100 provides a learning model for predicting the probability of occurrence of a negative event, which is an event that occurs within a predetermined period by a store opened on a commercial transaction site on a network and is a negative event for the commercial transaction site. Learn (step S201). Then, the information processing apparatus 100 determines the conditions for purchasing the receivables of the target store based on the probability of occurrence of the negative event by the target store predicted by using the learning model (step S202).

Next, FIG. 18 will be described. For example, FIG. 18 shows an example of determination of conditions regarding the purchase of receivables according to the economic situation of the society performed by the information processing apparatus 100. In FIG. 18, the information processing apparatus 100 indicates the number of occurrences of negative events, which are events caused by a store opened on a commercial transaction site on the network and which are negative events for the commercial transaction site, and social situation information indicating the economic situation of the society. Is estimated using (step S301). Then, the information processing apparatus 100 determines the conditions for purchasing the receivables based on the estimated number of occurrences (step S302).

Next, FIG. 19 will be described. For example, FIG. 19 shows an example of determining the upper limit of purchase of receivables performed by the information processing apparatus 100. In FIG. 19, the information processing apparatus 100 acquires store sales information indicating the past sales of the target store, which is a store that opens on a commercial transaction site on the network (step S401). Further, the information processing apparatus 100 acquires category sales information indicating past sales for each category of products handled on the commercial transaction site (step S402). Then, the information processing apparatus 100 is based on the sales tendency of the target store indicated by the store sales information and the sales tendency of the category corresponding to the handled product which is the product handled by the target store indicated by the category sales information. The maximum purchase amount of the receivables is determined (step S403).

Next, FIG. 20 will be described. For example, FIG. 20 shows an example of processing using a plurality of models performed by the information processing apparatus 100. In FIG. 20, the information processing apparatus 100 acquires a first learning model that predicts the probability of occurrence of a negative event, which is an event caused by a store opened on a commercial transaction site on a network and is a negative event for the commercial transaction site (). Step S501). Further, the information processing apparatus 100 acquires a second learning model for predicting the future possibility of the sales of the store opened on the commercial transaction site (step S502). Then, the information processing apparatus 100 is a target based on the probability of occurrence of a negative event by the target store predicted by using the first learning model and the future possibility of sales of the target store predicted by using the second learning model. The conditions for purchasing the store's receivables are determined (step S503).

Next, FIG. 21 will be described. For example, FIG. 21 shows an example of determining a mode of collecting accounts receivable corresponding to future receivables performed by the information processing apparatus 100. In FIG. 21, the information processing apparatus 100 estimates the closing probability, which is the probability that the target store, which is a store opening on the commercial trading site on the network, will leave the commercial trading site after a predetermined period (step S601). Then, the information processing apparatus 100 determines the collection mode of the accounts receivable corresponding to the future receivables of the target store based on the closing probability of the target store (step S602).

Next, FIG. 22 will be described. For example, FIG. 22 shows an example of determination of a proposal destination for the purchase of receivables performed by the information processing apparatus 100. In FIG. 22, the information processing apparatus 100 acquires store information about each of the plurality of stores that open on the commercial transaction site on the network (step S701). Then, the information processing apparatus 100 determines, among the plurality of stores, the proposal destination store to be the proposal destination for the purchase of the receivables, based on the store information regarding each of the plurality of stores (step S702).

[4. effect〕
As described above, the information processing apparatus 100 according to the embodiment has a learning unit 132 and a determination unit 136. The learning unit 132 learns a learning model for predicting the probability of occurrence of a negative event, which is an event that occurs within a predetermined period by a store opened on a commercial transaction site on the network and is a negative event for the commercial transaction site. The determination unit 136 determines the conditions for purchasing the receivables of the target store based on the probability of occurrence of a negative event by the target store predicted by using the learning model learned by the learning unit 132.

As described above, the information processing apparatus 100 according to the embodiment determines the conditions for purchasing the receivables of the target store based on the probability of occurrence of the negative event by the target store predicted by using the learning model. Since it is possible to purchase the receivables according to the above, it is possible to provide an appropriate service regarding the purchase of the receivables.

Further, in the information processing apparatus 100 according to the embodiment, the learning unit 132 learns a learning model in which the store information of the store is input and a score indicating the probability of occurrence of a negative event caused by the store is output.

As described above, the information processing apparatus 100 according to the embodiment is appropriate for the purchase of receivables by learning a learning model that outputs a score indicating the probability that a negative event occurs due to the store by inputting the store information of the store. The service can be made available.

Further, in the information processing apparatus 100 according to the embodiment, the learning unit 132 learns a learning model for predicting the probability of occurrence of uncollectible store receivables.

As described above, the information processing apparatus 100 according to the embodiment can provide an appropriate service regarding the purchase of receivables by learning a learning model for predicting the probability that the receivables of the store will become uncollectible. ..

Further, in the information processing apparatus 100 according to the embodiment, the determination unit 136 determines the conditions that are advantageous to the target store as the probability that the receivables of the target store become uncollectible is low.

As described above, the information processing apparatus 100 according to the embodiment provides an appropriate service regarding the purchase of receivables by determining conditions that are advantageous to the target store as the probability that the receivables of the target store become uncollectible is low. Can be made possible.

Further, in the information processing apparatus 100 according to the embodiment, the learning unit 132 learns a learning model for predicting the probability of store exit from a commercial transaction site.

As described above, the information processing apparatus 100 according to the embodiment can provide an appropriate service regarding the purchase of receivables by learning a learning model for predicting the probability of store exit from a commercial transaction site. can.

Further, in the information processing apparatus 100 according to the embodiment, the determination unit 136 determines the conditions that are more advantageous to the target store as the probability of occurrence of the closing of the target store is lower.

As described above, the information processing apparatus 100 according to the embodiment makes it possible to provide an appropriate service regarding the purchase of receivables by determining conditions that are advantageous to the target store as the probability of occurrence of the closing of the target store is low. be able to.

Further, in the information processing apparatus 100 according to the embodiment, the determination unit 136 determines the conditions for purchasing future receivables of the target store based on the probability of occurrence of a negative event by the target store.

As described above, the information processing apparatus 100 according to the embodiment provides an appropriate service regarding the purchase of receivables by determining the conditions for the purchase of future receivables of the target store based on the probability of occurrence of a negative event by the target store. Can be made possible.

Further, in the information processing apparatus 100 according to the embodiment, the determination unit 136 determines the index value for the purchase of future receivables of the target store.

As described above, the information processing apparatus 100 according to the embodiment can provide an appropriate service regarding the purchase of receivables by determining the index value of the future purchase of receivables of the target store.

Further, in the information processing apparatus 100 according to the embodiment, the determination unit 136 determines the upper limit of the purchase of future receivables of the target store based on the index value of the target store and the sales of the target store.

As described above, the information processing apparatus 100 according to the embodiment provides an appropriate service regarding the purchase of receivables by determining the upper limit of the purchase of future receivables of the target store based on the index value and the sales of the target store. Can be made possible.

Further, in the information processing apparatus 100 according to the embodiment, the determination unit 136 determines the purchase upper limit amount of the target store based on the value obtained by multiplying the index value of the target store and the sales of the target store.

As described above, the information processing apparatus 100 according to the embodiment is appropriate for the purchase of receivables by determining the upper limit of the purchase of future receivables of the target store based on the value obtained by multiplying the index value of the target store and the sales. Service can be provided.

Further, in the information processing apparatus 100 according to the embodiment, the determination unit 136 determines the number of months for which the future receivables of the target store are to be purchased as an index value.

As described above, the information processing apparatus 100 according to the embodiment can provide an appropriate service regarding the purchase of receivables by determining the number of months for which the future receivables of the target store are to be purchased as an index value. ..

Further, in the information processing apparatus 100 according to the embodiment, the learning unit 132 is a plurality of learning models, each of which corresponds to a different period, and predicts the probability that a negative event will occur due to the store within the corresponding period. Learn multiple learning models. The determination unit 136 uses a plurality of learning models to determine the number of months for which future receivables of the target store are to be purchased as an index value.

As described above, the information processing apparatus 100 according to the embodiment uses a plurality of learning models corresponding to different periods to determine the number of months for which future receivables of the target store are to be purchased as an index value. It is possible to provide appropriate services regarding the purchase of receivables.

Further, in the information processing apparatus 100 according to the embodiment, the learning unit 132 learns a plurality of learning models that output scores indicating the probability of occurrence of a negative event by the store within the corresponding period. The determination unit 136 determines the number of months for the purchase of future receivables of the target store based on the period corresponding to the learning model having the shortest period among the learning models whose scores are equal to or higher than the threshold value.

As described above, the information processing apparatus 100 according to the embodiment uses a plurality of learning models, each of which corresponds to a different period, and corresponds to the learning model having the shortest period among the learning models having a score equal to or higher than the threshold value. By determining the number of months for which future receivables of the target store are to be purchased based on the period, it is possible to provide an appropriate service regarding the purchase of receivables.

Further, the information processing apparatus 100 according to the embodiment has a providing unit 137. The providing unit 137 provides information indicating the conditions determined by the determining unit 136.

As described above, the information processing apparatus 100 according to the embodiment can provide appropriate information by providing information indicating the conditions.

[5. Hardware configuration]
The information processing apparatus 100 according to the above-described embodiment is realized by, for example, a computer 1000 having a configuration as shown in FIG. 23. FIG. 23 is a hardware configuration diagram showing an example of a computer that realizes the functions of the information processing device. The computer 1000 has a CPU 1100, a RAM 1200, a ROM 1300, an HDD (Hard Disk Drive) 1400, a communication interface (I / F) 1500, an input / output interface (I / F) 1600, and a media interface (I / F) 1700.

The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400, and controls each part. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started, a program depending on the hardware of the computer 1000, and the like.

The HDD 1400 stores a program executed by the CPU 1100, data used by such a program, and the like. The communication interface 1500 receives data from another device via the network N and sends it to the CPU 1100, and transmits the data generated by the CPU 1100 to the other device via the network N.

The CPU 1100 controls an output device such as a display or a printer, and an input device such as a keyboard or a mouse via the input / output interface 1600. The CPU 1100 acquires data from the input device via the input / output interface 1600. Further, the CPU 1100 outputs the generated data to the output device via the input / output interface 1600.

The media interface 1700 reads a program or data stored in the recording medium 1800 and provides the program or data to the CPU 1100 via the RAM 1200. The CPU 1100 loads the program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable Disk), a magneto-optical recording medium such as MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. And so on.

For example, when the computer 1000 functions as the information processing apparatus 100 according to the embodiment, the CPU 1100 of the computer 1000 realizes the function of the control unit 130 by executing the program loaded on the RAM 1200. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from another device via the network N.

Although some of the embodiments and modifications of the present application have been described in detail with reference to the drawings, these are examples and various based on the knowledge of those skilled in the art, including the embodiments described in the disclosure line of the invention. It is possible to carry out the present invention in another form in which the above is modified or improved.

[6. others〕
Further, among the processes described in the above-described embodiments and modifications, all or part of the processes described as being automatically performed can be manually performed, or are described as being manually performed. It is also possible to automatically perform all or part of the processed processing by a known method. In addition, information including processing procedures, specific names, various data and parameters shown in the above documents and drawings can be arbitrarily changed unless otherwise specified. For example, the various information shown in each figure is not limited to the information shown in the figure.

Further, each component of each of the illustrated devices is a functional concept, and does not necessarily have to be physically configured as shown in the figure. That is, the specific form of distribution / integration of each device is not limited to the one shown in the figure, and all or part of them may be functionally or physically distributed / physically in any unit according to various loads and usage conditions. Can be integrated and configured.

Further, the above-described embodiments and modifications can be appropriately combined as long as the processing contents do not contradict each other.

Further, the above-mentioned "section, module, unit" can be read as "means" or "circuit". For example, the acquisition unit can be read as an acquisition means or an acquisition circuit.

1 Information processing system 100 Information processing device 121 Store owner information storage unit 122 Model information storage unit 123 Threshold information storage unit 124 Group information storage unit 125 Social situation information storage unit 126 Commercial transaction site information storage unit 127 Factoring information storage unit 130 Control unit 131 Acquisition part 132 Learning part 133 Estimating part 134 Classification part 135 Judgment part 136 Decision part 137 Providing part 10 User terminal 20 Store device 30 Web server N network

Claims (16)

A learning unit that learns a learning model that predicts the probability of occurrence of a negative event that occurs within a predetermined period by a store that opens on a commercial transaction site on the network and is a negative event for the commercial transaction site.
A decision unit that determines the conditions for purchasing the receivables of the target store based on the probability of occurrence of the negative event by the target store predicted by using the learning model learned by the learning unit.
Equipped with
The learning unit
A plurality of learning models, each corresponding to a different period, are trained to predict the probability of occurrence of the negative event caused by the store within the corresponding period.
The decision-making part
An information processing apparatus comprising the plurality of learning models to determine the conditions regarding the period for which the receivables of the target store are to be purchased . The learning unit
The information processing apparatus according to claim 1, wherein the information processing apparatus according to claim 1 uses the store information of the store as an input and learns the learning model that outputs a score indicating the probability that the negative event occurs due to the store. The learning unit
The information processing apparatus according to claim 1 or 2, wherein the learning model for predicting the probability of occurrence of the claim of the store becoming uncollectible is learned. The decision-making part
The information processing apparatus according to claim 3, wherein the lower the probability that the claim of the target store becomes uncollectible, the more favorable the conditions for the target store are determined. The learning unit
The information processing apparatus according to any one of claims 1 to 4, wherein the learning model for predicting the probability of store exit from the commercial transaction site is learned. The decision-making part
The information processing apparatus according to claim 5, wherein the lower the probability of the occurrence of the exit of the target store, the more favorable the conditions for the target store are determined. The decision-making part
The information processing apparatus according to any one of claims 1 to 6, wherein the conditions for purchasing future receivables of the target store are determined based on the probability of occurrence of the negative event by the target store. The decision-making part
The information processing apparatus according to claim 7, wherein the index value for the purchase of the future claim of the target store is determined. The decision-making part
The information processing apparatus according to claim 8, wherein the upper limit of purchase of the future receivables of the target store is determined based on the index value of the target store and the sales of the target store. The decision-making part
The information processing apparatus according to claim 9, wherein the upper limit of purchase of the target store is determined based on a value obtained by multiplying the index value of the target store by the sales of the target store. The decision-making part
The information processing apparatus according to any one of claims 8 to 10, wherein the number of months for which the future claim of the target store is to be purchased is determined as the index value. The decision-making part
The information processing apparatus according to claim 11, wherein the number of months for which the future receivables of the target store are to be purchased is determined as the index value by using the plurality of learning models. The learning unit
Each trains the plurality of learning models that output a score indicating the probability of occurrence of the negative event caused by the store within the corresponding period.
The decision-making part
A claim characterized in that the number of months for which the future bond of the target store is to be purchased is determined based on the period corresponding to the learning model having the shortest period among the learning models whose score is equal to or higher than the threshold value. Item 12. The information processing apparatus according to Item 12. A provider that provides information indicating the conditions determined by the determination unit,
The information processing apparatus according to any one of claims 1 to 13, further comprising. It is an information processing method executed by a computer.
A learning process for learning a learning model that predicts the probability of occurrence of a negative event, which is an event that occurs within a predetermined period due to a store opened on a commercial transaction site on the network and is a negative event for the commercial transaction site.
A determination step of determining conditions for purchasing receivables of the target store based on the probability of occurrence of the negative event by the target store predicted using the learning model learned by the learning process.
Including
The learning process is
A plurality of learning models, each corresponding to a different period, are trained to predict the probability of occurrence of the negative event caused by the store within the corresponding period.
The determination step is
Using the plurality of learning models, the conditions regarding the period for which the receivables of the target store are to be purchased are determined.
An information processing method characterized by that. A learning procedure for learning a learning model for predicting the probability of occurrence of a negative event, which is an event that occurs within a predetermined period by a store opened on a commercial transaction site on a network and is a negative event for the commercial transaction site.
A decision procedure for determining the conditions for purchasing the receivables of the target store based on the probability of occurrence of the negative event by the target store predicted using the learning model learned by the learning procedure.
Let the computer run
The learning procedure is
A plurality of learning models, each corresponding to a different period, are trained to predict the probability of occurrence of the negative event caused by the store within the corresponding period.
The decision procedure is
An information processing program comprising the plurality of learning models to determine the conditions regarding the period for which the receivables of the target store are to be purchased .

Images