JP7195187B2 - Estimation device, estimation method and estimation program - Google Patents

Description

The present invention relates to an estimating device, an estimating method, and an estimating program.

Conventionally, techniques for predicting various demands have been provided. For example, in a management system that manages information registered by multiple businesses, a business that owns information of a type that is useful for the purpose of use received from users and that is not registered as a management target A technique of identifying a business operator and requesting registration from the identified business operator is known. Also, there is known a technique of estimating effects by using and not using registered information and providing rewards based on the estimated effects.

Japanese Patent No. 6370454

However, it may not be possible to easily provide rewards in an appropriate manner with the conventional techniques described above.

For example, in the conventional technology described above, the effect is calculated for each of cases in which registered information is used and not used, and the effect of the information is estimated based on the calculated effect. For this reason, the calculation cost increases, and it cannot be said that rewards in an appropriate manner can be easily provided.

In addition, in the above-described conventional technology, the effect of information selected by a data scientist or the like according to the purpose of use is only estimated. could not be automatically estimated.

The present application has been made in view of the above, and aims at an estimation device, an estimation method, and an estimation program that can easily provide rewards in an appropriate manner.

An estimating device according to the present application includes an identifying unit that identifies correlations among multiple types of information, and a prediction unit that predicts a predetermined type of information from the multiple types of information based on the correlations estimated by the estimating unit. and an estimating unit for estimating the degree of contribution to model accuracy.

According to one aspect of the embodiment, there is an effect that an appropriate aspect of remuneration can be easily provided.

FIG. 1 is a diagram illustrating an example of estimation processing performed by an estimation device according to an embodiment. FIG. 2 is a diagram illustrating an example of identification processing for identifying a degree of correlation performed by the estimation device according to the embodiment; FIG. 3 is a diagram illustrating a configuration example of an estimation system according to the embodiment; FIG. 4 is a diagram illustrating a configuration example of an estimation device according to the embodiment; FIG. 5 is a diagram illustrating an example of a registration information database according to the embodiment; FIG. 6 is a flowchart illustrating an example of the flow of estimation processing executed by the estimation device according to the embodiment. FIG. 7 is a hardware configuration diagram showing an example of a computer that implements the functions of the estimation device.

Embodiments for implementing the estimation device, estimation method, and estimation program according to the present application (hereinafter referred to as "embodiments") will be described in detail below with reference to the drawings. Note that the estimation device, estimation method, and estimation program according to the present application are not limited by this embodiment. Further, each embodiment can be appropriately combined within a range that does not contradict the processing contents. Also, in each of the following embodiments, the same parts are denoted by the same reference numerals, and overlapping descriptions are omitted.

[1. Example of Estimation Process Shown by Estimation Apparatus]
An example of estimation processing performed by the estimation device 100 according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of estimation processing performed by the estimation device 100 according to the embodiment. Specifically, the estimating apparatus 100 estimates the degree of contribution to the accuracy of a prediction model that predicts information of a predetermined type from information of a plurality of types, based on the correlation of the information of the plurality of types.

It should be noted that the business operator referred to here may be any business operator, and is a business operator that has previously formed an alliance with respect to exchange of information in a predetermined service. For example, business operators include business operators that operate restaurants, business operators that conduct retail business, business operators that operate fitness stores, business operators that provide public transportation, and mobile operators such as taxis. business operators, etc. that provide services related to

An example of estimation processing by the estimation device 100 will be described below along the flow with reference to FIG.

First, as shown in FIG. 1, the estimating device 100 receives data individually owned by each business operator (step S1). For example, in the example of FIG. 1, the estimation device 100 receives cafe visit data from business operator #1, general store visit data from business operator #2, and gym visit data from business operator #3.

Subsequently, the estimating device 100 receives a request for a model desired by the operator #X (step S2). For example, in the example of FIG. 1, the estimating device 100 receives a request for a prediction model regarding visits to a cafe as a model desired by business operator #X.

Then, the estimating apparatus 100 identifies the degree of correlation between the history of the data to be predicted and the data registered by other business operators (step S3). That is, the estimating apparatus 100 determines whether or not the user continuously uses different services by identifying the degree of correlation between the history of the data to be predicted and the data registered by other business operators. Information can be identified.

Specifically, estimating apparatus 100 identifies the correlation that information of multiple types has. In addition, specifically, the estimation device 100 identifies correlations of information indicating demands for different targets. More specifically, the estimation device 100 identifies the correlation according to the degree of similarity between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type. Also, more specifically, the number of users who visited a predetermined store or the number of users who purchased a predetermined transaction target is identified.

More specifically, the estimating apparatus 100 determines whether or not the temporal shift between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type satisfies a predetermined condition. based on which correlations are identified. More specifically, the estimation device 100 determines that the transition of demand indicated by the information of the first type is similar to the transition of demand indicated by the information of the predetermined type, and the demand indicated by the information of the first type is similar to the transition of demand indicated by the information of the first type. If the change occurs after the change in demand indicated by the predetermined type of information, the correlation between the first type of information and the predetermined type of information is made higher than a predetermined value.

For example, in the example of FIG. 1, estimation apparatus 100 identifies the degree of correlation between the cafe visit data and the general store visit data as correlation degree #1. Estimation apparatus 100 also identifies the degree of correlation between the cafe visit data and the gym visit data as correlation degree #2.

Here, the identification processing for identifying the degree of correlation between data will be further described with reference to FIG. FIG. 2 is a diagram illustrating an example of identification processing for identifying a degree of correlation performed by the estimation device according to the embodiment; An example of the specifying process shown in FIG. 2 is the process corresponding to step S3 in FIG. As shown in FIG. 2, FIG. FG1 is a diagram showing the relationship between the number of visitors and time.

For example, as shown in FIG. 2, FIG. FG1 shows cafe visit data regarding a user's visit to a cafe and gym visit data regarding a user's visit to a gym. Further, for example, in the example of FIG. 2, the estimation device 100 identifies the degree of correlation between the cafe visit data and the gym visit data by a conventional technique such as a calculation method for calculating the degree of correlation by time series analysis. For example, in the example of FIG. 2, the estimation apparatus 100 determines that the correlation is high because the linearity of the cafe visit data and the linearity of the gym visit data are similar to each other. . That is, the estimation apparatus 100 can identify the service that the user is expected to use next by identifying the degree of correlation between the cafe visit data and the gym visit data.

Returning to FIG. 1, an example of estimation processing according to the embodiment will be described. The estimating apparatus 100 generates a model for predicting visits to a cafe using data with a high degree of correlation (step S4). That is, estimating apparatus 100 identifies data with a high degree of correlation in advance, and uses the identified data to generate a model related to prediction of coming to a cafe, so that a highly accurate model can be generated. Specifically, the estimating apparatus 100 generates, as a prediction model, a prediction model that predicts information of one type among the plurality of types as information of a predetermined type from information of a plurality of types.

For example, in the example of FIG. 1, the estimation device 100 uses conventional machine learning techniques such as logistic regression, SVM (Support Vector Machine), and deep learning to generate data highly correlated with cafe visit data, such as gym visit data and , bus usage data, etc., and learning the relationship between the degree of correlation and the model for predicting café visits. That is, estimation apparatus 100 learns a model so that when specific data is input, the degree of correlation between the input specific data and cafe visit data is output.

In this way, in the example of FIG. 1 , estimation apparatus 100 identifies gym visit data and bus usage data as data highly correlated with cafe visit data among various data. is input, the model is trained to output the degree of correlation between the input specific data and the cafe visit data. In addition, the estimating apparatus 100 obtains data related to a visit by a user to a specific store and data related to a visit by a user to a different store from statistical data including data related to other users as well as data related to the same user. to determine the degree of correlation with data on store visits. Then, the estimating apparatus 100 uses conventional machine learning techniques such as logistic regression, SVM, and deep learning to generate input information, which is data highly correlated with data related to the visit of a user to a specific store, and the degree of correlation. By learning the relevance to the store, we generate a model for store visit prediction that users will visit a specific store.

Subsequently, the estimation device 100 provides the learned prediction model (step S5). For example, in the example of FIG. 1, the estimating device 100 provides the business operator #X with the generated model related to prediction of coming to a cafe.

Then, the estimation device 100 estimates the degree of contribution of each type of information based on the degree of correlation, and determines a reward according to the estimated degree of contribution (step S6). That is, the estimating apparatus 100 estimates the degree of contribution of each business operator's data to the model, and provides each business operator with a remuneration based on the estimated contribution degree, so that fair remuneration can be provided.

Specifically, the estimation apparatus 100 estimates the degree of contribution to the accuracy of a prediction model that predicts a predetermined type of information from multiple types of information based on the correlation. Specifically, the estimating apparatus 100 estimates the degree of contribution to the accuracy of a prediction model that predicts, from multiple types of information, one of multiple types of information as a predetermined type of information as a prediction model. do.

More specifically, estimation apparatus 100 identifies the correlation between each type of information and a predetermined type of information, and estimates the degree of contribution of a higher value as the identified correlation is higher. Also, specifically, estimation apparatus 100 determines a reward for a registrant who has registered each type of information based on the estimated degree of contribution.

For example, the estimating apparatus 100 estimates the degree of contribution of each business operator as a score of 0 to 1 as the degree of contribution between data received from each business operator based on the degree of correlation between data received from each business operator. For example, the estimating apparatus 100 estimates the contribution of a higher value as the correlation is higher. Also, for example, the estimation apparatus 100 estimates the contribution of a lower value as the correlation is lower. Then, the estimating device 100 determines a reward corresponding to each business operator, which is a reward according to the degree of contribution.

For example, in the example of FIG. 1, the estimation device 100 estimates the contribution #1 from the correlation #1 between the cafe visit data and the general store visit data. Then, the estimating apparatus 100 determines reward #1 corresponding to degree of contribution #1 as reward for business operator #1 who provided the cafe visit data and business operator #2 who provided the general store visit data.

Also, for example, in the example of FIG. 1, the estimation device 100 estimates the contribution #2 from the correlation #2 between the cafe visit data and the gym visit data. Then, the estimating apparatus 100 determines reward #2 corresponding to degree of contribution #2 as the reward for business operator #1 who provided the cafe visit data and business operator #3 who provided the gym visit data.

In this way, the estimation apparatus 100 according to the embodiment identifies correlations among multiple types of information. Based on this correlation, the estimation apparatus 100 estimates the degree of contribution to the accuracy of a prediction model that predicts a predetermined type of information from multiple types of information. As a result, the estimation apparatus 100 according to the embodiment estimates the degree of contribution to the accuracy of a prediction model that predicts a predetermined type of information from multiple types of information based on the correlation of the multiple types of information. rewards can be easily provided.

This point will be explained. To explain using the example of FIG. 1, the estimation device 100 estimates the degree of contribution of each business operator as the degree of contribution between data received from each business operator based on the degree of correlation between data received from each business operator. Then, the estimating device 100 determines a reward corresponding to each business operator, which is a reward according to the degree of contribution. As a result, the estimating apparatus 100 estimates the degree of contribution of each business operator's data to the model, and thus can determine a fair reward according to the degree of contribution. As a result, the estimating device 100 can provide fair rewards, so that the threshold for providing data can be lowered for businesses. In addition, since the estimating apparatus 100 can acquire useful data from the business operator, it is possible to improve the prediction accuracy of the model related to user usage prediction and the like. Therefore, the estimating apparatus 100 can acquire various useful data by providing a fair reward as a better incentive for the business operator, which makes it easier to formulate a comprehensive business strategy.

[2. Configuration of estimation system]
Next, the configuration of the estimation system according to the embodiment will be described using FIG. FIG. 3 is a diagram illustrating a configuration example of an estimation system according to the embodiment; As shown in FIG. 3 , estimation system 1 includes business operator server 20 , payment server 30 , and estimation device 100 . The operator server 20, the settlement server 30, and the estimation device 100 are connected via a network N so as to be able to communicate by wire or wirelessly. Note that the estimation system 1 shown in FIG. 3 may include multiple operator servers 20 , multiple payment servers 30 , and multiple estimation devices 100 .

The provider server 20 according to the embodiment is an information processing device used by each provider, and is realized by, for example, a server device, a cloud system, or the like. For example, the operator server used by each operator is denoted as operator server 20-N (N is an integer and is a value corresponding to N in operator #N). In this case, in the example of FIG. 1, the business server 20-1 is an information processing device used by the business operator #1 that operates the café. Further, the business server 20-2 is an information processing device used by the business operator #2 that operates the general store. Further, the business server 20-3 is an information processing device used by the business operator #3 that operates the gym.

Here, the business server 20 manages information about services provided by each business. Here, the information about the services provided by each business operator means, for example, information that associates the users who received the services provided by each business operator with the contents of the services provided to the users. It is the log of the service by the operator. In addition to such information, the provider server 20 can adopt arbitrary information as long as it is information related to services by each provider.

The payment server 30 according to the embodiment is an information processing device that performs payment processing, and is implemented by, for example, a server device. For example, when the estimation device 100 determines various rewards, the settlement server 30 cooperates with the estimation device 100 to execute settlement processing. In addition, for example, the settlement server 30 executes settlement processing by receiving notification from the estimation device 100 of the reward amount according to the degree of contribution of each type estimated based on the degree of correlation. Note that the settlement server 30 may perform settlement processing using, for example, a credit card, electronic settlement, or the like, in addition to settlement processing using cash.

The estimation device 100 according to the embodiment is realized by, for example, a server device. Specifically, estimating apparatus 100 identifies the correlation that information of multiple types has. Based on this correlation, the estimation apparatus 100 estimates the degree of contribution to the accuracy of a prediction model that predicts a predetermined type of information from multiple types of information.

[3. Configuration of estimation device]
Next, the configuration of the estimation device 100 according to the embodiment will be described using FIG. FIG. 4 is a diagram illustrating a configuration example of the estimation device 100 according to the embodiment. As shown in FIG. 4 , estimation device 100 includes communication unit 110 , storage unit 120 , and control unit 130 .

(Regarding communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card) or the like. The communication unit 110 is connected to a network by wire or wirelessly, and transmits and receives information between the business server 20 and the settlement server 30 .

(Regarding storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or flash memory, or a storage device such as a hard disk or an optical disk. The storage unit 120 has a registration information database 121 and a model database 122 .

(Regarding registration information database 121)
The registration information database 121 according to the embodiment stores information registered by each business operator, that is, information registered in the data management service. Here, FIG. 5 shows an example of the registration information database 121 according to the embodiment. In the example shown in FIG. 5, information having items such as "enterprise ID", "data ID", "subset ID", "user ID", "gender", "purchase history", etc. is associated and stored. ing. Further, as shown in FIG. 5, the information registered by each business is individually managed in the registration information database 121 without changing the format from the time of registration. In addition to the items shown in FIG. 5, the registration information database 121 may store information in which arbitrary items set by each business operator are set, and the value of each item is set by each business operator. may have been

For example, a “company ID” is an identifier that indicates a company that has registered information. "Data ID" is an identifier for identifying information registered by a business. For example, when a business operator registers the first information and then registers the second information, individual data IDs are assigned to the first information and the second information. A "subset ID" is an identifier that indicates a subset to which data belongs, and is information that can be arbitrarily set by a provider. The operator ID, data ID, and subset ID are information automatically assigned when the operator registers data, and the information registered by the operator, that is, the information acquired by the operator, For example, it corresponds to a set of information such as "user ID", "gender", and "purchase history".

For example, "user ID" is the identifier of the user who received the service of the business indicated by the business ID. "Gender" is the sex of the user indicated by the associated "user ID", and "purchase history" is information indicating the purchase history of the user.

Here, in the example shown in FIG. 5, the registration information database 121 stores the operator ID "operator A", the data ID "DID#1", the subset ID "subset #1", and the user ID "user#1". , gender “null”, and purchase history “purchase history #1” are registered in association with each other. Such information is information registered by the operator indicated by the operator ID "operator A", and belongs to the subset indicated by the subset ID "subset #1" in the data ID "DID#1". Indicates that the record is registered. In addition, this record has items such as "user ID", "sex", and "purchase history", and indicates that values such as "user #1", "null", and "purchase history #1" are stored. indicates

(Regarding the control unit 130)
The control unit 130 is a controller, and for example, various programs (an example of an estimation program) stored in a storage device inside the estimation device 100 are controlled by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. equivalent) is implemented by executing the RAM as a work area. Also, the control unit 130 is a controller, and is implemented by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

As shown in FIG. 4, the control unit 130 includes a receiving unit 131, a specifying unit 132, an estimating unit 133, a generating unit 134, and a determining unit 135, and functions and actions of information processing described below. realize or perform Note that the internal configuration of the control unit 130 is not limited to the configuration shown in FIG. 4, and may be another configuration as long as it performs information processing to be described later. Moreover, the connection relationship of each processing unit of the control unit 130 is not limited to the connection relationship shown in FIG. 4, and may be another connection relationship.

(Regarding the reception unit 131)
The reception unit 131 receives various information. Specifically, the reception unit 131 receives data owned individually by each business operator. For example, in the example of FIG. 1, the estimation device 100 receives cafe visit data from business operator #1, general store visit data from business operator #2, and gym visit data from business operator #3. Then, the reception unit 131 stores the data individually owned by each business in the registration information database 121 .

Further, specifically, the reception unit 131 receives a request for a model desired by a business operator. For example, in the example of FIG. 1, the reception unit 131 receives a request for a prediction model regarding visits to a cafe as a model desired by business operator #X.

(Regarding the specifying unit 132)
The identifying unit 132 identifies correlations among multiple types of information. Specifically, the identifying unit 132 identifies correlations of information indicating demands for different targets. More specifically, the identifying unit 132 identifies the correlation according to the degree of similarity between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type. More specifically, the correlation between the number of users who visited a predetermined store or the number of users who purchased a predetermined transaction object is specified.

Further, more specifically, the identifying unit 132 determines whether or not the temporal lag between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type satisfies a predetermined condition. based on which correlations are identified. More specifically, the identification unit 132 determines that the transition of demand indicated by the information of the first type is similar to the transition of demand indicated by the information of the predetermined type, and the demand indicated by the information of the first type is similar to the transition of demand indicated by the information of the predetermined type. If the change occurs after the change in demand indicated by the predetermined type of information, the correlation between the first type of information and the predetermined type of information is made higher than a predetermined value.

For example, in the example of FIG. 1, the identifying unit 132 identifies the degree of correlation between the cafe visit data and the general store visit data as the correlation degree #1. Further, the identifying unit 132 identifies the degree of correlation between the cafe visit data and the gym visit data as a correlation degree #2.

For example, in the example of FIG. 2, the specifying unit 132 specifies the degree of correlation between the cafe visit data and the gym visit data by a conventional technique such as a calculation method for calculating the degree of correlation by time series analysis. For example, in the example of FIG. 2, the identification unit 132 determines that the correlation is high because the linearity of the cafe visit data and the linearity of the gym visit data are similar to each other. .

(Regarding the estimation unit 133)
Based on the correlation specified by the specifying unit 132, the estimation unit 133 estimates the degree of contribution to the accuracy of a prediction model that predicts a predetermined type of information from multiple types of information. Specifically, the estimating unit 133 estimates, as a prediction model, the degree of contribution to the accuracy of a prediction model that predicts information of one type among a plurality of types as information of a predetermined type from information of a plurality of types.

More specifically, the estimation unit 133 identifies the correlation between each type of information and a predetermined type of information, and estimates the contribution of a higher value as the identified correlation is higher. For example, the estimation unit 133 estimates the degree of contribution of each business as the degree of contribution between data received from each business, based on the degree of correlation between data received from each business.

For example, the estimating unit 133 estimates the degree of contribution of each business as a score of 0 to 1 as the degree of contribution between data received from each business, based on the degree of correlation between data received from each business. For example, the estimation unit 133 estimates the contribution of a higher value as the correlation is higher. Also, for example, the estimating unit 133 estimates the contribution of a lower value as the correlation is lower. For example, in the example of FIG. 1, the estimation unit 133 estimates the contribution #1 from the correlation #1 between the cafe visit data and the general store visit data. Also, for example, the estimation unit 133 estimates the contribution #2 from the correlation #2 between the cafe visit data and the gym visit data.

(Regarding the generation unit 134)
The generation unit 134 generates a model for predicting a prediction target using data with a high degree of correlation. Specifically, the generation unit 134 generates, as a prediction model, a prediction model for predicting information of one type among the plurality of types as information of a predetermined type from information of a plurality of types.

For example, in the example of FIG. 1, the generation unit 134 uses conventional machine learning techniques such as logistic regression, SVM, and deep learning to generate data highly correlated with cafe visit data, such as gym visit data and bus usage data. By learning the relationship between the input information and the degree of correlation, a model for predicting cafe visits is generated. In this way, the generation unit 134 performs model learning so that when specific data is input, the degree of correlation between the input specific data and the cafe visit data is output. The generator 134 then stores the model in the model database 122 .

In addition, the generation unit 134 provides the learned prediction model to the business operator by referring to the model database 122 . For example, in the example of FIG. 1, the generating unit 134 provides the generated model related to prediction of coming to a cafe to the operator #X.

(About decision unit 135)
Based on the degree of contribution estimated by the estimation unit 133, the determination unit 135 determines a reward for a registrant who has registered each type of information. For example, the determination unit 135 determines a reward corresponding to each business operator, which is a reward according to the degree of contribution. Then, the determination unit 135 cooperates with the settlement server 30 to set the remuneration to be paid to the business operator.

For example, in the example of FIG. 1, the determination unit 135 assigns a reward #1 corresponding to the degree of contribution #1 to the operator #1 who provided the cafe visit data and the operator #2 who provided the general store visit data. Reward and decide.

Further, for example, in the example of FIG. 1, the determining unit 135 assigns a reward #2 corresponding to the degree of contribution #2 to the operator #1 who provided the cafe visit data and the operator #3 who provided the gym visit data. Reward and decide.

[4. Processing procedure]
Next, with reference to FIG. 6, a procedure of estimation processing executed by the estimation device 100 according to the embodiment will be described. FIG. 6 is a flowchart showing an example of the flow of estimation processing executed by the estimation device 100 according to the embodiment.

As shown in FIG. 6, the reception unit 131 receives a model request (step S101). For example, in the example of FIG. 1, the reception unit 131 receives a request for a prediction model regarding visits to a cafe as a model desired by business operator #X.

If the accepting unit 131 has not accepted a model request (step S101; No), the specifying unit 132 waits until the accepting unit 131 accepts a model request.

On the other hand, when the receiving unit 131 receives a model request (step S101; Yes), the identifying unit 132 identifies the history of the data to be estimated (step S102). The specifying unit 132 then specifies the correlation between the specified data and other data (step S103).

For example, in the example of FIG. 1, the identifying unit 132 identifies the degree of correlation between the cafe visit data and the general store visit data as the correlation degree #1. Further, the identifying unit 132 identifies the degree of correlation between the cafe visit data and the gym visit data as a correlation degree #2.

For example, in the example of FIG. 2, the specifying unit 132 specifies the degree of correlation between the cafe visit data and the gym visit data by a conventional technique such as a calculation method for calculating the degree of correlation by time series analysis. For example, in the example of FIG. 2, the identification unit 132 determines that the correlation is high because the linearity of the cafe visit data and the linearity of the gym visit data are similar to each other. .

Then, the estimation unit 133 estimates the degree of contribution to the estimation accuracy based on the correlation (step S104). For example, the estimating unit 133 estimates the degree of contribution of each business as a score of 0 to 1 as the degree of contribution between data received from each business, based on the degree of correlation between data received from each business. For example, the estimation unit 133 estimates the contribution of a higher value as the correlation is higher. Also, for example, the estimating unit 133 estimates the contribution of a lower value as the correlation is lower.

For example, in the example of FIG. 1, the estimation unit 133 estimates the contribution #1 from the correlation #1 between the cafe visit data and the general store visit data. Also, for example, the estimation unit 133 estimates the contribution #2 from the correlation #2 between the cafe visit data and the gym visit data.

Then, the generating unit 134 generates a model using data whose contribution degree estimated by the estimating unit 133 exceeds a predetermined threshold, and provides the model (step S105). For example, in the example of FIG. 1, the generation unit 134 uses conventional machine learning techniques such as logistic regression, SVM, and deep learning to generate data highly correlated with cafe visit data, such as gym visit data and bus usage data. By learning the relationship between the input information and the degree of correlation, a model for predicting cafe visits is generated. In this way, the generation unit 134 performs model learning so that when specific data is input, the degree of correlation between the input specific data and the cafe visit data is output.

In addition, the generation unit 134 provides the learned prediction model to the business operator by referring to the model database 122 . For example, in the example of FIG. 1, the generating unit 134 provides the generated model related to prediction of coming to a cafe to the operator #X.

Then, the determining unit 135 determines a reward according to the degree of contribution (step S106). For example, the determination unit 135 determines a reward corresponding to each business operator, which is a reward according to the degree of contribution. Then, the determination unit 135 sets a reward by cooperating with the settlement server 30 .

For example, in the example of FIG. 1, the determination unit 135 assigns a reward #1 corresponding to the degree of contribution #1 to the operator #1 who provided the cafe visit data and the operator #2 who provided the general store visit data. Reward and decide.

Further, for example, in the example of FIG. 1, the determining unit 135 assigns a reward #2 corresponding to the degree of contribution #2 to the operator #1 who provided the cafe visit data and the operator #3 who provided the gym visit data. Reward and decide.

[5. Modification]
The estimation device 100 described above may be implemented in various different forms other than the above embodiment. Therefore, other embodiments of the estimation device 100 will be described below.

[5-1. Specific Processing User Attribute]
In the above-described embodiment, an example of the identification process in which the identification 132 of the estimating device 100 identifies the correlation according to the degree of similarity between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type has been described. However, it is not limited to the above specific processing. Specifically, the identifying unit 132 may identify the correlation that the attributes of the user have.

For example, assume that the data used to identify the degree of correlation are cafe visit data and gym visit data. In this case, the specifying unit 132 uses a conventional technique such as a calculation method for calculating the degree of correlation by time-series analysis to determine the correlation between the user attributes included in the cafe visit data and the user attributes included in the gym visit data. Determine degree.

As a result, the identifying unit 132 of the estimation device 100 according to the embodiment can identify information about users who continuously use services provided by each business operator in order to identify the correlation of attributes of users. can be done.

[5-2. Generation process Contribution-based model]
In the above-described embodiment, an example of generation processing in which the generation unit 134 of the estimation device 100 generates a model for predicting a prediction target using data with a high degree of correlation has been described, but the generation processing is not limited to the above-described generation processing. Specifically, the generation unit 134 may estimate the degree of contribution from the degree of correlation of each type of information, and generate a model for automatically predicting a prediction target using information with a high degree of estimated contribution.

For example, the estimating unit 133 estimates the degree of contribution of each business as a score of 0 to 1 as the degree of contribution between data received from each business, based on the degree of correlation between data received from each business. For example, the estimation unit 133 estimates the contribution of a higher value as the correlation is higher. Also, for example, the estimating unit 133 estimates the contribution of a lower value as the correlation is lower. For example, assume that the degree of correlation between the data on visiting a cafe and the data on visiting a general store is #1. In this case, the estimation unit 133 estimates the contribution from the correlation #1 between the cafe visit data and the general store visit data.

Then, the generation unit 134 generates a model for predicting a prediction target using data with a high degree of contribution. For example, it is assumed that data with a high degree of contribution are cafe visit data, gym visit data, and bus usage data. In this case, the generation unit 134 generates input information such as cafe visit data, gym visit data, and bus usage data as data with a high degree of contribution, using conventional machine learning techniques such as logistic regression, SVM, and deep learning. A model may be generated by learning associations with contributions. In this way, the generation unit 134 performs model learning so that when specific data is input, the degree of contribution of the input specific data is output.

As a result, the generation unit 134 of the estimation device 100 according to the embodiment estimates the degree of contribution from the degree of correlation of each type of information, and automatically generates a model for predicting a prediction target using information with a high degree of estimated contribution. Since it is generated, appropriate forms of rewards can be easily provided.

[6. Hardware configuration]
Also, the business operator server 20, the settlement server 30, and the estimation device 100 according to the above-described embodiments are implemented by a computer 1000 configured as shown in FIG. 7, for example. Hereinafter, the estimation device 100 will be described as an example. FIG. 7 is a hardware configuration diagram showing an example of a computer 1000 that implements the functions of the estimation device 100. As shown in FIG. The computer 1000 includes a CPU 1100, a RAM 1200, a ROM (Read Only Memory) 1300, a 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 programs stored in the ROM 1300 or HDD 1400 and controls each section. The ROM 1300 stores a boot program executed by the CPU 1100 when the computer 1000 is started up, a program depending on the hardware of the computer 1000, and the like.

The HDD 1400 stores programs executed by the CPU 1100, data used by the programs, and the like. Communication interface 1500 receives data from other devices via network N, sends the data to CPU 1100, and transmits data generated by CPU 1100 via network N to other devices.

The CPU 1100 controls output devices such as displays and printers, and input devices such as keyboards and mice, through an input/output interface 1600 . CPU 1100 acquires data from an input device via input/output interface 1600 . Also, CPU 1100 outputs the generated data to an output device via input/output interface 1600 .

Media interface 1700 reads programs or data stored in recording medium 1800 and provides them to CPU 1100 via RAM 1200 . CPU 1100 loads such a program from recording medium 1800 onto RAM 1200 via 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 a PD (Phase change rewritable disc), a magneto-optical recording medium such as an MO (Magneto-Optical disk), a tape medium, a magnetic recording medium, or a semiconductor memory. etc.

For example, when the computer 1000 functions as the estimation device 100 according to the embodiment, the CPU 1100 of the computer 1000 implements the functions of the control unit 130 by executing programs loaded on the RAM 1200 . In addition, data in storage unit 120 is stored in HDD 1400 . 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 via the network N from another device.

[7. others〕
Further, among the processes described in the above embodiments and modifications, all or part of the processes described as being performed automatically can be performed manually, or described as being performed manually. All or part of the processing can also be performed automatically by known methods. 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 drawing is not limited to the illustrated information.

Also, each component of each device illustrated is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution and integration of each device is not limited to the one shown in the figure, and all or part of them can be functionally or physically distributed and integrated in arbitrary units according to various loads and usage conditions. Can be integrated and configured.

Also, the above-described embodiments and modifications can be appropriately combined within a range that does not contradict the processing contents.

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

[8. effect〕
As described above, the estimation device 100 according to the embodiment has the identification unit 132 and the estimation unit 133 . The identifying unit 132 identifies correlations among multiple types of information. Based on the correlation specified by the specifying unit 132, the estimation unit 133 estimates the degree of contribution to the accuracy of a prediction model that predicts a predetermined type of information from multiple types of information.

As a result, the estimation apparatus 100 according to the embodiment estimates the degree of contribution to the accuracy of a prediction model that predicts a predetermined type of information from a plurality of types of information based on the specified correlation. Rewards can be easily provided.

The estimation apparatus 100 according to the embodiment also has a determination unit 135 that determines a reward for a registrant who has registered each type of information based on the degree of contribution estimated by the estimation unit 133 .

As a result, the estimation apparatus 100 according to the embodiment determines a reward for a registrant who has registered each type of information based on the estimated degree of contribution, so it is possible to easily provide an appropriate form of reward. .

In addition, in the estimation device 100 according to the embodiment, the estimation unit 133, as a prediction model, predicts one type of information among a plurality of types of information as a predetermined type of information from a plurality of types of information. Estimate the contribution.

As a result, the estimation apparatus 100 according to the embodiment estimates the degree of contribution to the accuracy of a prediction model that predicts information of one of a plurality of types as a predetermined type of information from a plurality of types of information as a prediction model. Therefore, appropriate forms of remuneration can be easily provided.

Further, in the estimation apparatus 100 according to the embodiment, the identifying unit 132 identifies the correlation between each type of information and the predetermined type of information, and the estimating unit 133 determines that the correlation identified by the identifying unit 132 is Higher estimates the contribution of higher values.

As a result, the estimation apparatus 100 according to the embodiment estimates the degree of contribution of a higher value as the correlation with information of a predetermined type is higher for each type of information, so that an appropriate form of reward can be easily provided. can do.

In addition, in the estimation device 100 according to the embodiment, the identifying unit 132 identifies correlations of information indicating demands for different targets.

As a result, the estimation apparatus 100 according to the embodiment identifies the correlation of the information indicating demands for different targets, so that it is possible to easily provide an appropriate form of remuneration.

In addition, in the estimation device 100 according to the embodiment, the identifying unit 132 identifies the correlation according to the degree of similarity between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type.

As a result, the estimation device 100 according to the embodiment identifies the correlation according to the degree of similarity between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type. can be easily provided.

In addition, in the estimation device 100 according to the embodiment, the identification unit 132 determines whether the timing difference between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type satisfies a predetermined condition. Determine the correlation based on whether or not.

With this, the estimation device 100 according to the embodiment is based on whether or not the temporal lag between the transition of demand indicated by the information of each type and the transition of demand indicated by the information of the predetermined type satisfies a predetermined condition. to identify correlations, so that appropriate forms of rewards can be readily provided.

Further, in the estimation device 100 according to the embodiment, the identification unit 132 determines that the transition of demand indicated by the information of the first type is similar to the transition of demand indicated by the information of the predetermined type, and the information of the first type is If the indicated transition of demand occurs after the transition of demand indicated by the predetermined type of information, the correlation between the first type of information and the predetermined type of information is made higher than a predetermined value.

As a result, the estimation device 100 according to the embodiment determines that the transition of demand indicated by the information of the first type is similar to the transition of demand indicated by the information of the predetermined type, and the transition of demand indicated by the information of the first type is similar to the transition of demand indicated by the information of the predetermined type. However, if it occurs after the change in demand indicated by the information of the predetermined type, the correlation between the information of the first type and the information of the predetermined type is made higher than the predetermined value. Rewards can be easily provided.

In addition, in the estimation device 100 according to the embodiment, the identifying unit 132 identifies the correlation between the number of users who visited a predetermined store or the number of users who purchased a predetermined transaction target.

As a result, the estimation device 100 according to the embodiment determines the correlation between the number of users who have visited a predetermined store or the number of users who have purchased a predetermined transaction target, so that an appropriate form of remuneration is provided. can be easily provided.

In addition, in the estimation device 100 according to the embodiment, the identifying unit 132 further identifies correlations of attributes of the user.

As a result, the estimation device 100 according to the embodiment further identifies the correlation of the attributes of the user, so that it is possible to easily provide an appropriate form of remuneration.

As described above, some of the embodiments of the present application have been described in detail based on the drawings. It is possible to carry out the invention in other forms with modifications.

1 estimation system 20 operator server 30 settlement server 100 estimation device 110 communication unit 120 storage unit 121 registration information database 122 model database 130 control unit 131 reception unit 132 identification unit 133 estimation unit 134 generation unit 135 determination unit

Claims (7)

a specifying unit that specifies the correlation between the first type of information and the second type of information as the correlation of the plurality of types of information;
Contribution to the accuracy of a prediction model that predicts the first type of information based on the first type of information and the second type of information whose correlation with the first type of information is equal to or greater than a predetermined threshold an estimation unit for estimating the degree of
a determination unit that determines a reward for a registrant who has registered each type of information based on the degree of contribution estimated by the estimation unit;
has
The identification unit
According to the degree of similarity between the transition of demand indicated by the second type of information and the transition of demand indicated by the first type of information, the correlation possessed by the information indicating the demand for each different target is specified as the correlation,
The estimation unit
As the prediction model, the contribution to the accuracy of the prediction model that predicts the information of the first type, and the higher the correlation specified by the specifying unit, the higher the contribution of the value is estimated.
An estimation device characterized by: The identification unit
Identifying the correlation based on whether the temporal shift between the transition of demand indicated by the second type of information and the transition of demand indicated by the first type of information satisfies a predetermined condition,
The estimation device according to claim 1 , characterized by: The identification unit
The transition of demand indicated by the information of the second type is similar to the transition of demand indicated by the information of the first type, and the transition of demand indicated by the information of the second type is similar to the transition of demand indicated by the information of the first type If it occurs after the change in demand shown, the correlation between the second type information and the first type information is made higher than a predetermined value,
3. The estimating device according to claim 1 or 2 , characterized in that: The identification unit
As the correlation, specifying the correlation of the number of users who visited a predetermined store or the number of users who purchased a predetermined transaction target,
The estimation device according to any one of claims 1 to 3 , characterized in that: The identification unit
As the correlation, further specifying the correlation possessed by the attribute of the user,
5. The estimation device according to claim 4 , characterized in that: An estimation method executed by an estimation device,
an identifying step of identifying the correlation between the first type of information and the second type of information as the correlation of the plurality of types of information;
Contribution to the accuracy of a prediction model that predicts the first type of information based on the first type of information and the second type of information whose correlation with the first type of information is equal to or greater than a predetermined threshold an estimation step of estimating the degree of
a determining step of determining a reward for a registrant who has registered each type of information based on the contribution estimated by the estimating step;
including
The specific step includes
According to the degree of similarity between the transition of demand indicated by the second type of information and the transition of demand indicated by the first type of information, the correlation possessed by the information indicating the demand for each different target is specified as the correlation,
The estimation step includes
As the prediction model, it is a contribution to the accuracy of the prediction model that predicts the first type of information, and the higher the correlation identified by the identification step, the higher the contribution of the value is estimated.
An estimation method characterized by: an identifying procedure for identifying a correlation between information of a first type and information of a second type as correlations possessed by information of a plurality of types;
Contribution to the accuracy of a prediction model that predicts the first type of information based on the first type of information and the second type of information whose correlation with the first type of information is equal to or greater than a predetermined threshold an estimation procedure for estimating the degree of
a determination procedure for determining a reward for a registrant who has registered each type of information based on the contribution estimated by the estimation procedure;
on the computer, and
The specific procedure includes:
According to the degree of similarity between the transition of demand indicated by the second type of information and the transition of demand indicated by the first type of information, the correlation possessed by the information indicating the demand for each different target is specified as the correlation,
The estimation procedure includes:
As the prediction model, it is the contribution to the accuracy of the prediction model that predicts the first type of information, and the higher the correlation identified by the identification procedure, the higher the contribution of the value is estimated.
estimation program.

Images