JP7249103B2 - Selection device, selection method and selection program - Google Patents

Description

The present invention relates to a selection device, selection method, and selection program.

Conventionally, various technologies related to business in business operators (stores, etc.) have been provided. For example, there is provided a technique for performing sales management by predicting sales of merchandise in a store. In such a technique, for example, alarm information is output when the predicted product sales are not within a predetermined deviation allowable range from the sales target value.

JP 2004-127017 A

However, the conventional technology described above does not always enable appropriate actions for a given business operator. For example, if the predicted product sales are not within a predetermined deviation tolerance range from the sales target value, what kind of action should the side (for example, the store) that receives the output of the alarm information take to cope with the output of the alarm information? It may be difficult to enable appropriate actions for a given business operator due to the need to make a decision as to whether it should be taken.

The present application has been made in view of the above, and an object thereof is to provide a selection device, a selection method, and a selection program that enable appropriate action for a predetermined business operator.

A selection device according to the present application includes an acquisition unit that acquires a planned value of an index related to the business of one business operator that corresponds to a predetermined point in time and history information before the predetermined point in time; a prediction unit that predicts a predicted value of the index related to the business of the one business operator corresponding to the predetermined point in time based on the history information obtained by the obtaining unit; a selection unit that selects a coping action that is an action for coping with the deviation when the deviation between the predicted value predicted by the unit and the planned value satisfies a predetermined condition. .

According to one aspect of the embodiment, there is an effect that it is possible to enable appropriate action for a predetermined business operator.

FIG. 1 is a diagram illustrating an example of generation processing according to the embodiment. FIG. 2 is a diagram illustrating an example of selection processing according to the embodiment. FIG. 3 is a diagram illustrating a configuration example of a selection system according to the embodiment; FIG. 4 is a diagram illustrating a configuration example of a selection device according to the embodiment. 5 is a diagram illustrating an example of a learning data storage unit according to the embodiment; FIG. 6 is a diagram illustrating an example of a model information storage unit according to the embodiment; FIG. 7 is a diagram illustrating an example of a store information storage unit according to the embodiment; FIG. 8 is a diagram illustrating an example of an index information storage unit according to the embodiment; FIG. 9 is a diagram illustrating an example of a coping action information storage unit according to the embodiment; FIG. FIG. 10 is a flowchart illustrating an example of generation processing according to the embodiment. FIG. 11 is a flowchart illustrating an example of selection processing according to the embodiment. FIG. 12 is a hardware configuration diagram showing an example of a computer that implements the functions of the selection device.

Modes for implementing a selection device, a selection method, and a selection program (hereinafter referred to as "embodiments") according to the present application will be described in detail below with reference to the drawings. Note that the selection device, selection method, and selection program according to the present application are not limited to this embodiment. Also, in each of the following embodiments, the same parts are denoted by the same reference numerals, and overlapping descriptions are omitted.

(embodiment)
[1. Generation process]
First, an example of generation processing according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of generation processing according to the embodiment. In FIG. 1, the selection device 100 uses history information related to each of a plurality of businesses including one business to generate a model for predicting the predicted value of an index related to the business of one business. indicate the case. It should be noted that the selection device 100 may generate a model using only history information related to one business operator instead of history information related to each of a plurality of business operators. Moreover, any information may be included in the history information. For example, history information may include any information as long as it is information that can be acquired prior to the time of prediction. For example, the history information includes various types of information as long as it is information that can be used as input information for the model. Details of the history information will be described later. Also, FIG. 1 shows a case where each of the businesses has one store, that is, a case where the business and the store have a one-to-one relationship. For example, a case where the selection device 100 selects store A as a target store and generates a model for predicting the sales of store A is shown. Note that the example of FIG. 1 shows a case of predicting store sales as an example of store KPIs (Key Performance Indicators). You can create a model that predicts any metric. Also, the selection device 100 may target a case where one business operator has a plurality of stores, such as chain stores. That is, the selection device 100 may generate a model for predicting the predicted value of the index related to the business of one business operator using history information related to one business operator having multiple stores. In this case, the selection device 100 may use the model to predict an indicator related to the performance of a business having multiple stores. For example, the selection device 100 generates a model (model MX) using only history information of stores located in a predetermined area (such as area X) among stores of one business operator having a plurality of stores. may In this case, the selection device 100 may use the model MX to predict indicators related to the performance of the area XX of one operator. As described above, the business operator targeted by the selection device 100 is not limited to a business operator having one store, and may be various business operators such as a business operator having a plurality of stores. In the following, for the sake of simplification of explanation, the description will be based on the case where one business operator owns one store, so the business operator will be described as a store. That is, hereinafter, the store may be read as the business operator.

In FIG. 1, the selection device 100 inputs various information such as history information of a plurality of stores such as store A and store B (hereinafter also referred to as "store history information"), weather information, and user behavior information. The case of using as (input data) is shown. Below, the history information associated with the correct answer information is also referred to as “learning data”. Further, when the date and time are described below, it may be a predetermined period (for example, 10 minutes, 1 hour, 1 day, etc.) including the date and time.

Further, in FIG. 2, the selection device 100 uses the generated model to obtain store history information such as store A and store B at a certain date and time (period), user behavior information at a certain date and time (period), and store A store history information. Based on history information including weather information and the like corresponding to the area, the sales of store A at a predetermined date and time (period) after that date and time are predicted. Hereinafter, the predicted sales value is also referred to as a sales prediction value. In FIG. 2, the selection device 100 performs coping actions based on the discrepancy between the sales forecast value of store A at a predetermined date and time and the sales value planned by store A (hereinafter also referred to as "sales plan value"). Indicates when to select.

[Configuration of selection system]
Prior to the description of FIGS. 1 and 2, the configuration of the selection system 1 will be described with reference to FIG. FIG. 3 is a diagram illustrating a configuration example of a selection system according to the embodiment; As shown in FIG. 3 , the selection system 1 includes a terminal device 10 , a store terminal 20 , a service providing device 50 and a selection device 100 . The terminal device 10, the store terminal 20, the service providing device 50, and the selection device 100 are connected via a predetermined network N so as to be communicable by wire or wirelessly. The selection system 1 shown in FIG. 3 may include a plurality of terminal devices 10, a plurality of store terminals 20, a plurality of service providing devices 50, and a plurality of selection devices 100. .

The terminal device 10 is an information processing device used by a user. The terminal device 10 is realized by, for example, a smart phone, a tablet terminal, a notebook PC (Personal Computer), a desktop PC, a mobile phone, a PDA (Personal Digital Assistant), or the like. FIG. 1 shows a case where the terminal device 10 is a smart phone. Also, the terminal device 10 has a function such as a GPS (Global Positioning System) sensor, etc., and is capable of detecting and acquiring the position of the user. In addition, the terminal device 10 transmits various information and the like to the selection device 100 and the service providing device 50 . For example, the terminal device 10 transmits location information about the user and the like to the selection device 100 and the service providing device 50 . The terminal device 10 also receives information from the selection device 100 and the service providing device 50 . For example, the terminal device 10 receives information from the selection device 100 and the service providing device 50 . In addition, below, the terminal device 10 may be described as a user. That is, hereinafter, the user can also be read as the terminal device 10 .

The terminal device 10 may collect user information such as user behavior information and transmit it to the selection device 100 and the service providing device 50 . Moreover, the terminal device 10 may detect various sensor information by various sensors other than a GPS sensor or the like. For example, the terminal device 10 may predict and acquire the user's location information using the location information of the base station with which it is communicating and the radio waves of WiFi (registered trademark) (Wireless Fidelity). The terminal device 10 may also have a function of an acceleration sensor, and may detect and acquire acceleration information (sensor information) in the movement of the user. Further, the terminal device 10 may have various functions such as a temperature sensor and an air pressure sensor, and may be capable of detecting and acquiring environment information such as temperature and air pressure in which the user is placed. In addition, the terminal device 10 may have various functions such as a heartbeat sensor, and may be capable of detecting and acquiring biological information of the user. For example, a user using the terminal device 10 may acquire the user's own context information from the terminal device 10 by wearing a wearable device capable of communicating with the terminal device 10 . For example, a user using the terminal device 10 wears a wristband-type wearable device that can communicate with the terminal device 10, so that the terminal device 10 acquires information about the user's own heartbeat (pulse). It may be possible.

The store terminal 20 is an information processing device used by each store. For example, the store terminal 20 is used by the manager of each store. The store terminal 20 transmits and receives information to and from the terminal device 10 , the selection device 100 and the service providing device 50 . For example, the manager of store A uses the store terminal 20 to transmit and receive information to and from the terminal device 10 , the selection device 100 and the service providing device 50 .

In addition, the store terminal 20 transmits various information and the like to the selection device 100 and the service providing device 50 . For example, the shop terminal 20 transmits information about the shop to the selection device 100 and the service providing device 50 . The store terminal 20 also receives information from the selection device 100 and the service providing device 50 . For example, the shop terminal 20 receives information from the selection device 100 and the service providing device 50 . For example, the store terminal 20 is realized by a smart phone, a tablet terminal, a notebook PC, a desktop PC, a mobile phone, a PDA, or the like. In addition, below, the store terminal 20 may be described as a store. That is, hereinafter, the store can also be read as the store terminal 20 .

The service providing device 50 is an information processing device used by service providers. For example, the service providing device 50 is used by a service provider who distributes content to users. The service providing device 50 transmits and receives information to and from the terminal device 10 , the shop terminal 20 and the selection device 100 . In addition, the service providing device 50 transmits various information and the like to the terminal device 10 , the store terminal 20 and the selection device 100 . For example, the service providing device 50 transmits information about the service to be provided to the terminal device 10, the store terminal 20, and the selection device 100. FIG. The service providing device 50 also receives information from the terminal device 10 , the store terminal 20 and the selection device 100 . For example, the service providing device 50 receives information from the terminal device 10, the store terminal 20, or the selection device 100. FIG.

The service providing device 50 provides various information to the selecting device 100 . For example, the service providing device 50 provides the selecting device 100 with collected weather information and information on the user. For example, the service providing device 50 provides the selecting device 100 with weather information indicating the weather of the area A and the like. For example, the service providing device 50 transmits user location information and user number information indicating the number of users located in area A to the selecting device 100 . For example, the service providing device 50 provides the selection device 100 with information on various user behaviors such as user searches, purchases, and content browsing. Further, for example, the service providing device 50 may provide model information such as the model M1 to the selecting device 100 .

The selection device 100 is an information processing device that selects a coping action, which is an action for coping with the deviation, when the deviation between the predicted value and the planned value for the index related to the business of one store satisfies a predetermined condition. is. The selection device 100 predicts the predicted value of the index corresponding to a predetermined point in time based on history information related to each of the plurality of shops before the predetermined point in time. Further, when a coping action is selected, the selection device 100 executes processing corresponding to the selected coping action. For example, when a coping action is selected, the selection device 100 provides a service such as providing information to the terminal device 10, the store terminal 20, or the service providing device 50 based on the selected coping action. For example, the selection device 100 transmits information to the terminal device 10, the store terminal 20, and the service providing device 50. FIG. For example, the selection device 100 transmits information based on coping behaviors to the terminal device 10, the store terminal 20, and the service providing device 50. FIG.

First, with reference to FIG. 1, a case where the selection device 100 generates a model M1 used for predicting the sales of store A will be described. In the example of FIG. 1, it is assumed that information indicating the sales plan of store A (hereinafter also referred to as "sales plan information") has already been obtained. For example, the selection device 100 acquires sales plan information including sales plan values CV10 to CV16 corresponding to dates DA10 to DA16, as shown in the index information storage unit 124 (see FIG. 8). For example, the selection device 100 acquires sales plan information from the store terminal 20 of store A (see FIG. 3) and stores it in the index information storage unit 124 (see FIG. 8). In addition, in the example of FIG. 1, a case where A store located in area A is targeted as shown in map information MP1 will be described as an example.

In addition, "* (* is an arbitrary number)" following "DA" in each date and time "DA*" shown in the examples of FIGS. make new. For example, the date and time 'DA11' has a smaller number following 'DA' than the date and time 'DA12'. . For example, each date and time "DA*" may be a predetermined period such as one hour.

The selection device 100 acquires history information of each store from a plurality of stores. The selection device 100 acquires history information related to store KPIs for each date and time (period). For example, the selection device 100 acquires history information regarding sales, the number of visitors, the number of inquiries, etc. of store A at each date and time (period).

The selection device 100 acquires history information of store A from store A (step S11-1). The selection device 100 acquires history information of store A from the store terminal 20 of store SH11 (store A) in the map information MP1 of FIG. For example, the selection device 100 acquires store history information LG11 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 1, the selection device 100 acquires store history information LG11-0 to LG11-3 corresponding to dates and times DA10 to DA13 from A store. Note that, as described above, the date and time referred to here may be a predetermined period including the date and time. For example, the date and time may be a predetermined period before that date and time (for example, within 24 hours from that date and time). For example, date and time DA11 and date and time DA12 may include overlapping periods.

The selection device 100 also acquires history information of the B store from the B store (step S11-2). The selection device 100 acquires the history information of the B store from the store terminal 20 of the store SH12 (B store) in the map information MP1 of FIG. For example, the selection device 100 acquires store history information LG12 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 1, the selection device 100 acquires store history information LG12-0 to LG12-3 corresponding to the dates and times DA10 to DA13 from the B store.

Hereinafter, steps S11-1 and S11-2 will be collectively referred to as step S11 when described without distinguishing between them. Information acquisition from each store may be performed multiple times, not limited to steps S11-1 and S11-2. Note that FIG. 1 shows a case where information on only two stores, the A store and the B store, is used for the sake of simplicity of explanation, but the selection device 100 uses the information on the stores other than the A store to be predicted. You may use not only the information of a store but many stores (For example, 100 stores, 1000 stores, etc.).

The selection device 100 also acquires correct information from the A store (step S12). The selection device 100 acquires the correct answer information from the shop terminal 20 of the A shop. For example, the selection device 100 acquires correct information RDT1 indicating the actual sales value (hereinafter also referred to as "measured value") corresponding to each date and time from the A store. In the example of FIG. 1, the selection device 100 acquires the correct information RDT1-1 to RDT1-4 corresponding to the dates and times DA11 to DA14 from the A store.

The selection device 100 also acquires history information including weather information and information on user behavior (hereinafter also referred to as “user-related information”) from the service providing device 50 (step S13). For example, the selection device 100 acquires history information LG21 corresponding to each date and time from the service providing device 50 . In the example of FIG. 1, the selection device 100 acquires history information LG21 including weather information and user-related information corresponding to dates DA10 to DA13 from the service providing device 50. In the example of FIG.

For example, the selection device 100 acquires weather information related to the location of store A as weather information. For example, the selection device 100 acquires weather information of area A corresponding to each date and time as weather information related to the location of store A.

For example, the selection device 100 acquires a user's action history on the Internet as user-related information. For example, the selection device 100 acquires information such as search information indicating a user's search and purchase information indicating a user's purchase as user-related information. The purchase information may be purchase information related to products sold by store A or products related to the products. Also, for example, the selection device 100 acquires information about the user's schedule and schedule, such as calendar information, as the user-related information. The selection device 100 also acquires, as user-related information, user number information indicating the number of users located in area A where A store is located. In the example of FIG. 1, only search information is illustrated as user-related information, but the history information LG21 also includes purchase information corresponding to each date and time, information on user schedules, and user number information. . Further, steps S11 to S13 may be performed at any timing as long as it is before step S14. For example, step S13 may be performed before step S11 or step S12.

Then, the selecting device 100 adds data obtained by combining the information obtained in steps S11 to S13 as learning data (step S14). In the example of FIG. 1, the selection device 100 generates learning data using the actual sales values (measured sales values) for each date and time DA11 to DA14 included in the correct information RDT1 as the correct information RDT1-1 to RDT1-4. . Specifically, the selection device 100 adds learning data to the learning data storage unit 121 .

For example, the selection device 100 generates data DT1-1 including correct information RDT1-1 indicating the actual sales value of store A on date and time DA11 as learning data. The selection device 100 also generates data DT1-1 including store history information of stores A and B prior to the date and time DA11 as learning data. In the example of FIG. 1, the selection device 100 generates data DT1-1 including store history information LG11-0 of store A and store history information LG12-0 of store B corresponding to date and time DA10 as learning data. Further, the selection device 100 generates data DT1-1 including weather information and user-related information prior to the date and time DA11 as learning data. In the example of FIG. 1, the selection device 100 generates data DT1-1 including, as learning data, weather information WDT10 and search information SDT10 at a date and time DA10 earlier than the date and time DA11, among the history information LG21.

For example, the search information SDT10 includes information indicating the number of searches using queries related to products provided by store A on date and time DA10. For example, the search information SDT10 includes information indicating the number of searches using the names of products provided by store A on date and time DA10. Weather information WDT10 also includes information indicating whether the area (area A) associated with store A at date and time DA10 was sunny, cloudy, rainy, or the like. The data DT1-1 includes the purchase information of the user at the date and time DA10 and the number of users in the area A as user-related information. Also, the weather information WDT10 may be information predicted at the time of data addition (for example, at step S14). The weather information WDT10 may be, for example, weather information indicating whether the weather at the date and time DA11 will be sunny, cloudy, rainy, or the like.

As described above, when "data DT* (* is an arbitrary numerical value)" is described, it indicates that the data is identified by the data ID "DT*". For example, when "data DT1-1" is written, the data is identified by the data ID "DT1-1".

Further, for example, the selection device 100 generates data DT1-2 including correct information RDT1-2 indicating the actual sales value of store A on date and time DA12 as learning data. The selection device 100 also generates data DT1-2 including store history information of store A and store B prior to date and time DA12 as learning data. In the example of FIG. 1, the selection device 100 generates data DT1-2 including store history information LG11-1 of store A and store history information LG12-1 of store B corresponding to date and time DA11 as learning data. Further, the selection device 100 generates data DT1-2 including weather information and user-related information prior to the date and time DA12 as learning data. In the example of FIG. 1, the selection device 100 generates, as learning data, data DT1-2 including, as history information, weather information WDT11 and search information SDT11 at date and time DA11 earlier than date and time DA12, among history information LG21.

For example, the search information SDT11 includes information indicating the number of searches using queries related to products provided by store A on date and time DA11. For example, the search information SDT11 includes information indicating the number of searches using the names of products provided by store A on date and time DA11. Weather information WDT11 also includes information indicating whether the area (area A) associated with store A at date and time DA11 was sunny, cloudy, rainy, or the like. The data DT1-2 includes the purchase information of the user at the date and time DA11 and the number of users in the area A as user-related information. Also, the weather information WDT11 may be information predicted at the time of data addition (for example, at the time of step S14). The weather information WDT11 may be, for example, weather information indicating whether the weather at the date and time DA12 will be sunny, cloudy, rainy, or the like.

Further, the correct information RDT1-1 to RDT1-4 are not limited to actual sales values, and may be various information based on actual sales values. For example, it may be information indicating to which level an actual sales value for a predetermined period (for example, one day) of store A corresponds. For example, the correct information RDT1-1 to RDT1-4 may be 10 levels from 0 to 1 if the maximum daily sales of store A is 100,000 yen. For example, the correct information RDT1-1 to RDT1-4 are level "0" when the measured value is less than 10,000 yen, and level "0. 1”. In addition, the correct information RDT1-1 to RDT1-4 has a level of "0.9" when the measured value is 90,000 yen or more and less than 100,000 yen, and a level of "0.9" when the measured value is 100,000 yen or more. 1”. In this manner, the correct information RDT1-1 to RDT1-4 may be information indicating which level among a plurality of levels corresponds, that is, information indicating sales in stages. For example, the correct information RDT1-1 may be "0.7" when the actual sales value on the date and time DA11 is 70,000 yen.

In this case, the selection device 100 may store a correspondence table indicating correspondence between each level and sales in the storage unit 120 (see FIG. 4). The model M1 may be a model that outputs a score indicating a level. The selection device 100 may predict the predicted value of sales using the score output by the model M1 and the level correspondence table. For example, if the model M1 to which history information is input outputs a score of "0.9", the selection device 100 may predict the sales to be "90,000 yen" using the level correspondence table. . To simplify the explanation, the case where the selection device 100 generates the model M1 that outputs values between 0 and 1 will be explained below as an example.

The selection device 100 generates a model based on the learning data (step S15). In this way, the selection device 100 generates a model using store history information, weather information, and user-related information at a date and time before the date and time corresponding to each correct answer information as history information. As a result, the selection device 100 can generate a model for predicting future sales of store A by inputting current information.

For example, the selection device 100 generates a model by performing learning using the data DT1-1 to DT1-4 and the like in the learning data storage unit 121 as learning data (teacher data). For example, the selection device 100 generates a model using feature amounts related to store history information. For example, the selection device 100 generates a model using feature amounts related to store KPIs. Also, for example, the selection device 100 generates a model using feature amounts related to weather information. For example, the selection device 100 generates a model using feature values relating to the weather, temperature, humidity, etc. of an area related to a store. In addition, for example, the selection device 100 generates a model using feature amounts related to search information on stores. For example, the selection device 100 generates a model using feature amounts related to the number of searches using queries related to products provided by a store. In this way, the selection device 100 learns the features included in the history information and generates a model for predicting store sales. In the example of FIG. 1, the selection device 100 learns features related to information included in history information and generates a model for predicting store sales. The features included in the history information learned by the selection device 100 may be input to the selection device 100 by a person such as an administrator of the selection device 100, or may be automatically learned (extracted) by the selection device 100. good too.

For example, the selection device 100 uses learning data as shown in the learning data storage unit 121 to generate the model M1. For example, when the correct information RDT1-1 is "1" indicating that the sales were 100,000 yen or more, the selection device 100 selects when the history information included in the data DT1-1 is input to the model M1. , the learning process is performed so that the score output by the model M1 approaches "1". For example, when the store history information LG11-0, the store history information LG12-0, the weather information WDT10, and the search information SDT10 of the data DT1-1 are input to the model M1, the selection device 100 determines that the score output by the model M1 is Learning processing is performed so as to approach "1".

Further, for example, when the correct information RDT1-2 is "0.5" indicating that the sales were 50,000 yen, the selection device 100 inputs the history information included in the data DT1-2 to the model M1. In this case, the learning process is performed so that the score output by the model M1 approaches "0.5". For example, when the store history information LG11-1, the store history information LG12-1, the weather information WDT11, and the search information SDT11 of the data DT1-2 are input to the model M1, the selection device 100 determines that the score output by the model M1 is Learning processing is performed so as to approach "0.5".

Note that the model learning method is not limited to the method described above, and any known technique can be applied. Note that each model may be generated using various conventional techniques related to machine learning as appropriate. For example, a model may be generated using a technology related to supervised learning machine learning such as SVM (Support Vector Machine). Also, for example, model generation may be performed using techniques related to unsupervised machine learning. For example, model generation may be performed using deep learning techniques. For example, the model may be generated using various deep learning techniques such as DNN (Deep Neural Network), RNN (Recurrent Neural Network), and CNN (Convolutional Neural Network). Note that the above description of model generation is an example, and model generation may be performed by a learning method appropriately selected according to obtainable information and the like. That is, if the selection device 100 can learn the model M1 so as to output a score corresponding to the correct answer information when the history information included in the learning data is input, what kind of method is used to generate the model M1? may be performed.

By the above processing, the selection device 100 generates a model (model M1) identified by the model ID “M1” as shown in the model information storage unit 122 in the example of FIG. As described above, when "model M* (* is an arbitrary numerical value)" is described, it indicates that the model is identified by the model ID "M*". For example, when "model M1" is described, the model is identified by the model ID "M1". Further, as shown in the model information storage unit 122 in FIG. 1, the model M1 is a model used for use "sales forecast (A store)", that is, for sales forecast of A store, and its specific model data is "model data MDT1". For example, by inputting history information corresponding to a certain date and time in the model M1, the selection device 100 can store a score indicating the sales of store A ahead (future) of the date and time corresponding to the input history information to the model M1. The sales of store A are predicted based on the score output by the model M1.

As described above, the selection device 100 learns using the learning data in which the history information and the correct answer information are associated, thereby appropriately predicting the predicted value of the index related to the business of one store. A model can be generated. Therefore, by using the model generated as described above, the selection device 100 can, for example, accurately predict the predicted value of the index related to the business of one store. In the above example, the correct information RDT1-1 of the data DT1-1 is information indicating the sales of the A store on the date DA11. In addition, the data DT1-1 includes store history information LG11-0, store history information LG12-0, weather information WDT10, and search information SDT10 at a date and time DA10 past the date and time DA11 corresponding to the correct information RDT1-1 as history information. include. In this way, the selection device 100 generates a model using information before the date and time corresponding to the correct answer information RDT1-1 as input information. Accordingly, the selection device 100 can generate a model for predicting the predicted value of the index related to the business of one store by inputting the current situation.

[2. Selection process]
An example of selection processing according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram illustrating an example of selection processing according to the embodiment. FIG. 2 shows a case where the selection device 100 selects a coping action based on the discrepancy between the predicted sales value for date and time DA16 and the sales plan value for date and time DA16. Note that description of the same points as in FIG. 1 will be omitted as appropriate. It is assumed that the category of store A is "restaurant" as shown in the store information storage unit 123 in FIG.

The selection device 100 acquires history information of each store from a plurality of stores. The selection device 100 acquires the history information regarding the KPI of the store at the date and time DA15. For example, the selection device 100 acquires history information regarding the sales, the number of visitors, the number of inquiries, etc. of store A on date and time DA15.

The selection device 100 acquires history information of store A from store A (step S21-1). The selection device 100 acquires history information of store A from the store terminal 20 of store SH11 (store A) in the map information MP1 of FIG. For example, the selection device 100 acquires store history information LG11 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 2, the selection device 100 acquires store history information LG11-5 corresponding to date and time DA15 from A store.

Further, the selecting device 100 acquires history information of the B store from the B store (step S21-2). The selection device 100 acquires the history information of the B store from the store terminal 20 of the store SH12 (B store) in the map information MP1 of FIG. For example, the selection device 100 acquires store history information LG12 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 2, the selection device 100 acquires store history information LG12-5 corresponding to date and time DA15 from B store.

Hereinafter, steps S21-1 and S21-2 will be collectively referred to as step S21 when described without distinguishing between them. Information acquisition from each store may be performed multiple times, not limited to steps S21-1 and S21-2. Note that the selection device 100 may use information of a large number of stores (for example, 100 stores, 1000 stores, etc.) in addition to the information of the B store for stores other than the A store to be predicted.

The selecting device 100 also acquires history information including weather information and user-related information from the service providing device 50 (step S22). For example, the selecting device 100 acquires history information LG25 including weather information WDT15 and search information SDT15 corresponding to date and time DA15 from the service providing device 50 . For example, the selection device 100 acquires weather information related to the location of store A as weather information. For example, the selection device 100 acquires weather information of area A corresponding to each date and time as weather information related to the location of store A. Obtain weather information WDT15 and search information SDT15.

For example, the selection device 100 acquires information such as search information indicating a user's search and purchase information indicating a user's purchase as user-related information. The selection device 100 also acquires, as user-related information, user number information indicating the number of users located in area A where A store is located. Note that steps S21 and S22 may be performed at any timing before step S23. For example, step S22 may be performed before step S21.

The selection device 100 then generates input data by combining the information acquired in steps S21 and S22 (step S23). In the example of FIG. 2, the selection device 100 inputs data DT21 including store history information LG11-5 of store A and store history information LG12-5 of store B corresponding to date and time DA15, as shown in input data list IPD21. Generate as data. The selection device 100 also generates data DT21 including weather information and user-related information prior to the date and time DA16 as input data. In the example of FIG. 2, the selection device 100 generates data DT1-1 including weather information WDT15 and search information SDT15 as history information as learning data, as shown in the input data list IPD21.

Then, the selection device 100 inputs the data DT21 to the model. In the example of FIG. 2, the selection device 100 calculates the score indicating the sales of store A on date DA16 by performing the processing shown in processing group PS21. The selection device 100 inputs the data DT21 to the model M1 (step S24). Specifically, the selection device 100 inputs history information including store history information LG11-5, LG12-5, weather information WDT15, search information SDT15, etc. to the model M1. Model M1 to which data DT21 is input outputs a score (step S25). In the example of FIG. 2, model M1 to which data DT21 is input outputs score SC16. For example, the score SC16 may be a value indicating the sales level such as "0.2" or "0.3". The score SC16 may be a sales value such as "20000" or "30000".

Then, the selecting device 100 predicts the predicted sales value for the date and time DA16 based on the score indicating the sales of the A store at the date and time DA16 (step S26). For example, the selection device 100 predicts the predicted sales value for date/time DA16 using a correspondence table showing the correspondence between each level and sales. In the example of FIG. 2, the selection device 100 predicts the sales forecast value PV16 for the date and time DA16 using a correspondence table showing the correspondence between each level and sales, as shown in the forecast list PL11. For example, when the score SC16 is "0.2", the selection device 100 uses the correspondence table showing the correspondence between each level and the sales, and the sales of the A store predicted on the date and time DA16 is 20,000 yen (for example, The sales forecast value PV16) is predicted. The above is just an example, and the selection device 100 may use various information as appropriate to predict (calculate) the sales prediction value PV16.

Then, the selecting device 100 calculates the difference between the planned sales value and the predicted sales value for the date and time DA16 (step S27). In the example of FIG. 2, the selection device 100 detects from the index information storage unit 124 (see FIG. 8) that the planned sales value for store A on date DA16 is the planned sales value CV16, as shown in the plan list CL11. Get the information shown. Then, the selection device 100 calculates that the difference between the planned sales value CV16 and the predicted sales value PV16 is the difference DF11, as shown in the divergence list DL11. For example, the selection device 100 calculates the difference by comparing the planned sales value and the predicted sales value. For example, the selection device 100 calculates the difference DF11 by subtracting the predicted sales value PV16 from the planned sales value CV16. In this case, when the sales forecast value is smaller than the sales plan value, the difference is a positive value, and when the sales forecast value is larger than the sales plan value, the difference is a negative value. For example, when the difference is a positive value, the selection device 100 can determine that there is a high possibility that future sales will not reach the planned level. In the example of FIG. 2, it is assumed that the sales forecast value PV16 is smaller than the sales plan value CV16. For example, the sales forecast value PV16 may be 20,000 yen and the sales plan value CV16 may be 50,000 yen.

Then, the selecting device 100 determines whether or not the difference DF11 indicating the deviation between the planned sales value CV16 and the predicted sales value PV16 satisfies the condition (step S28). For example, the selection device 100 uses the conditions corresponding to the category “eating and drinking” and the index “sales” of store A among the information stored in the coping action storage unit 125 in FIG. 1, and the difference DF11 satisfies the conditions. determine whether For example, the selection device 100 determines which of the divergence conditions between the object “eating and drinking” and the index “sales” stored in the coping action storage unit 125 in FIG. 1 corresponds to the divergence condition. For example, in the coping action storage unit 125 in FIG. 1, the coping action (coping action AT1) identified by the coping action ID "AT1" indicates "advertisement distribution." Note that the coping action AT1 may be various actions such as increasing the bid amount of the advertisement, increasing the number of times the advertisement is distributed, or increasing the exposure of the advertisement of the A store.

In the example of FIG. 2, the selection device 100 determines that the difference DF11 is greater than the condition value SP2 and less than the condition value SP3 by comparing the difference DF11 with the condition values SP1 to SP3. In this case, the selection device 100 determines that the divergence between the planned sales value and the predicted sales value of the store A satisfies the divergence condition of the condition value SP2.

Then, the selection device 100 selects a coping action based on the deviation between the planned value and the predicted value (step S29). The selection device 100 selects one measure from a plurality of measures for improving the sales of the store A as a coping action according to the degree of divergence between the planned value and the predicted value. For example, the selection device 100 selects one coping action from coping actions AT1 to AT3 related to the improvement of the sales of the store A according to the degree of divergence between the planned value and the predicted value. In the example of FIG. 2, the selection device 100 selects the coping action AT2 corresponding to the condition value SP2 because the difference between the planned sales value and the predicted sales value of the store A satisfies the conditions for the difference of the condition value SP2.

The selecting device 100 then executes the selected coping action (step S30). In the example of FIG. 2, the selection device 100 requests the service providing device 50 to distribute coupons to users located in area A where A store is located.

The service providing device 50 requested by the selecting device 100 to distribute the coupon to the users located in the area A distributes the coupon of the store A to the terminal device 10 of each user according to the position of each user. (Step S31). In the example of FIG. 2, the service providing device 50 distributes the coupon of the A store to the terminal device 10 of the user U1 located in the area A. As a result, the selection system 1 can motivate the user U1 to purchase products at the A store, and eliminate the discrepancy in sales at the A store.

As described above, the selection device 100 predicts the sales of store A at the target date and time for prediction based on the history information prior to the target date and time for prediction. In the example of FIG. 2, the selection device 100 causes the model M1 to output a score regarding the sales of the A store by inputting history information to the model M1. Then, the selection device 100 predicts the sales of the A store based on the score output by the model M1. Then, the selection device 100 selects a coping action based on the difference between the predicted sales forecast value and the sales plan value. As a result, the selection device 100 can enable appropriate actions for a given business operator. In addition, since the selection device 100 determines whether or not deviation occurs based on the predicted information, it is possible to eliminate the time lag until the actual measurement value becomes clear. That is, the selection device 100 selects which action (coping action ) can be selected. Therefore, the selection device 100 can select which action (coping action) to take without waiting for the confirmation of the actual measurement value. Therefore, the selection device 100 can reduce the time lag and make it possible to act on a predetermined store at an appropriate timing. Note that when the predicted value exceeds the planned value, the selection device 100 selects a coping action that suppresses an increase in the value of the index, such as lowering the bid amount of the advertisement or reducing the number of advertisements distributed. By doing so, a corrective action may be selected such that the predicted value matches the planned value as closely as possible.

[2-1. Executing body]
In the above example, the selection device 100 makes a request to the service providing device 50 based on the selected coping action, and the service providing device 50 performs the coping action. you can go In the example of FIG. 2, the selection device 100 may distribute coupons to users located in area A. In the example of FIG. The selection device 100 may transmit the coupon of the A store to the terminal device 10 of the user U1 located in the area A. Thus, the coping action selected by the selection device 100 may be performed by the selection device 100 itself, or may be performed by a device other than the selection device 100 such as the service providing device 50 .

[2-2. Use of information between stores]
For example, the selection device 100 may perform selection processing based on the sales forecast value only for stores that have permitted provision of information on other stores. For example, the selection device 100 may acquire from each store information indicating whether or not each store is permitted to use its own store's information (provide information) for prediction of indices of other stores. For example, the selection device 100 may acquire from store A information indicating whether or not store A is permitted to use the information of store A for predicting indices of other stores such as store B and store C. . In this case, the selection device 100 may perform selection processing only for stores that have permitted provision of information to other stores (permitted stores). For example, the selection device 100 uses the information of the one permitted store and the information of the other permitted stores to calculate the predicted value of the index of the one permitted store for the one permitted store that is permitted to provide information to the other stores. may be predicted and selection processing may be performed. As a result, the selection device 100 encourages each store to provide information, enables information collection from a large number of authorized stores, and uses the information to improve prediction accuracy. Therefore, the selection device 100 can improve the quality of services provided.

In addition, in the example of FIG. 1, in the prediction for store A located in area A, store information of store B located in area A, weather information of area A, etc. are used as history information. However, history information is not limited to area A information, and information of various areas may be used. For example, the selection device 100 may predict the predicted value of store A located in area A using weather information of areas other than area A and store information of stores located in other areas. For example, the selection device 100 may predict the predicted value of store A located in area A using weather information of an area adjacent to area A (adjacent area) and store information of stores located in the adjacent area. . For example, when predicting the predicted value of a store (pharmacy PHA) that sells drugs, the selection device 100 is located in an area adjacent to the area where the pharmacy PH is located, and is located in an area that sells drugs (pharmacy PHB). Store information may be used. For example, when predicting sales of cold medicines at a drug store (pharmacy PHA), the selection device 100 may use cold medicine sales information of a pharmacy PHB located in an area adjacent to the area of the pharmacy PH. As a result, the selection apparatus 100 can detect that, for example, if cold medicines are selling well in an area (neighboring town) adjacent to the area where the prediction target store is located, there is a possibility that the prediction target store will also start selling them. can be reasonably predicted. Note that the above is just an example, and the selection device 100 may perform prediction using information on various areas.

[2-3. About selection of coping behavior (action)]
In the example of FIG. 1, to simplify the explanation, the selection device 100 compares the deviation (difference) between the planned value and the predicted value with the deviation (reference) associated with each coping action. A case of selecting an action is shown. However, the selection device 100 may select coping actions through various processes using information on the divergence (difference) between the planned value and the predicted value. For example, the selection device 100 may select a coping action using a model that receives information on the divergence (difference) between the planned value and the predicted value. For example, the selection device 100 may select coping actions using a model that receives as input information on the divergence (difference) between the planned value and the predicted value, weather information on the weather, and information on other elements. For example, when the selection device 100 is input with information on the divergence (difference) between the planned value and the predicted value, weather information on the weather, and information on other elements, the selection device 100 performs coping actions (actions) corresponding to the input information. ) may be used to select a coping action.

Further, when inputting information about the divergence (difference) between the planned value and the predicted value, weather information about the weather, and information about other elements, the selection device 100 outputs a score corresponding to the input information. A model may be used to select a coping action. For example, the selection device 100 may select coping behaviors using a list in which scores (reference values) and coping behaviors (actions) are associated with each other. For example, the selection device 100 may compare a score output by a model to which weather information about weather and information about other elements is input with a score (reference value) in a list, and select coping actions. For example, the selection device 100 may select a coping action associated with a score (reference value) corresponding to the score output by the model from the list. Further, the selection device 100 may update (generate) the model through reinforcement learning based on the selected and executed coping action and the subsequent change in the index value. In this way, the selection device 100 may refer to other elements such as the weather in addition to the degree of divergence to select coping actions (actions) that maximize the effect, or present them to the business operator. Note that the above is just an example, and the selection device 100 may select coping behaviors (actions) using various models and other information.

[3. Configuration of Selection Device]
Next, the configuration of the selection device 100 according to the embodiment will be described using FIG. FIG. 4 is a diagram illustrating a configuration example of a selection device according to the embodiment. As shown in FIG. 4 , the selection device 100 has a communication section 110 , a storage section 120 and a control section 130 . The selection device 100 has an input unit (for example, a keyboard, a mouse, etc.) that receives various operations from the administrator of the selection device 100, and a display unit (for example, a liquid crystal display, etc.) for displaying various information. may

(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 the network N by wire or wirelessly, and transmits and receives information to and from external devices such as the terminal device 10, the store terminal 20, and the service providing device 50.

(storage unit 120)
The storage unit 120 is realized by, for example, a semiconductor memory device such as a RAM (Random Access Memory) or a flash memory, or a storage device such as a hard disk or an optical disk. As shown in FIG. 4, the storage unit 120 according to the embodiment includes a learning data storage unit 121, a model information storage unit 122, a shop information storage unit 123, an index information storage unit 124, and a coping behavior information storage unit 125. and

(Learning data storage unit 121)
The learning data storage unit 121 according to the embodiment stores various information about learning data. 5 is a diagram illustrating an example of a learning data storage unit according to the embodiment; FIG. For example, the learning data storage unit 121 stores teacher data used for model generation. The learning data storage unit 121 shown in FIG. 5 includes items such as "data ID", "correct answer information", and "history information". The "history information" includes items such as "store A", "store B", "weather information", and "search information".

"Data ID" indicates identification information for identifying data. For example, the data identified by the data ID "DT1-1" corresponds to the data DT1-1 shown in the example of FIG.

"Correct information" indicates correct information corresponding to the data identified by the data ID. For example, "correct answer information" indicates the actual index value corresponding to the data identified by the data ID. For example, "correct information" indicates the actual sales value corresponding to the data identified by the data ID. FIG. 5 shows an example in which conceptual information such as "RDT1-1" is stored in the "correct answer information", but in reality, the actual sales at store A, which is one store, at a predetermined date and time (period) are measured. Information indicating a value or a file path name indicating its storage location is stored. For example, the “correct information” may be information indicating the level of the measured value of sales.

The "correct information" is not limited to the actual sales value, and may be various information based on the actual sales value. For example, it may be information indicating to which level an actual sales value for a predetermined period (for example, one day) of store A corresponds. For example, if the maximum daily sales of store A is 100,000 yen, the "correct information" may be 10 levels from 0 to 1. For example, the "correct answer information" is level "0" when the measured value is less than 10,000 yen, and level "0.1" when the measured value is 10,000 yen or more and less than 20,000 yen. good too. In addition, the "correct information" is level "0.9" when the measured value is 90,000 yen or more and less than 100,000 yen, and is level "1" when the measured value is 100,000 yen or more. good too. In this way, the "correct answer information" may be information indicating which level among a plurality of levels corresponds, that is, information indicating sales in stages. In this case, the selection device 100 may store in the storage unit 120 a correspondence table indicating correspondence between each level and sales. The model M1 may be a model that outputs a score indicating a level. The selection device 100 may predict the predicted value of sales using the score output by the model M1 and the level correspondence table. For example, if the model M1 to which history information is input outputs a score of "0.9", the selection device 100 may predict the sales to be "90,000 yen" using the level correspondence table. .

"A store" in the "history information" includes various history information related to A store at a predetermined date and time (period). For example, "A store" in "history information" includes information related to the business of A store at a predetermined date and time (period). For example, "A store" in "history information" includes information about A store at a predetermined date and time (period). For example, "A store" in "history information" includes information such as sales, number of visitors, number of inquiries, etc. of A store at a predetermined date and time (period). FIG. 5 shows an example in which conceptual information such as "LG11-0" is stored in "A store", but in reality, history information related to A store at a predetermined date and time (period), or its A file path name that indicates the storage location is stored.

"B store" in "history information" includes various history information related to B store at a predetermined date and time (period). For example, "B store" in "history information" includes information related to the business of B store at a predetermined date and time (period). For example, "B store" in the "history information" includes information related to KPI of B store at a predetermined date and time (period). For example, "B store" in "history information" includes information such as sales, number of visitors, and number of inquiries at B store at a predetermined date and time (period). FIG. 5 shows an example in which conceptual information such as "LG12-0" is stored in "B store", but in reality, history information related to B store at a predetermined date and time (period), or its A file path name that indicates the storage location is stored.

“Weather information” in “history information” indicates weather information of an area related to the location of store A at a predetermined date and time (period). For example, "weather information" in "history information" indicates whether the weather in area A where store A is located was "sunny", "cloudy", or "rainy" at a predetermined date and time (period). It may be information. FIG. 5 shows an example in which conceptual information such as "WDT10" is stored in "weather information", but in reality, the address and latitude and longitude information indicating the specific location of the business operator, or the storage location thereof The file path name that indicates is stored.

"Search information" in "history information" indicates search information included in the history information. For example, the "search information" in the "history information" may be information indicating the number of searches using queries related to products provided by store A in a date and time (period). FIG. 5 shows an example in which conceptual information such as "SDT10" is stored in "search information", but in reality, various information such as the number of searches for store A at a predetermined date and time (period), or A file path name indicating the storage location is stored.

For example, in the example shown in FIG. 5, the data (data DT1-1) identified by the data ID "DT1-1" includes store history information LG11-0 of store A and store history information LG12 of store B as history information. -0, weather information WDT10, search information SDT10, etc. are included. Data DT1-1 also indicates that correct answer information RDT1-1 is included. For example, the correct answer information RDT1-1 of the data DT1-1 is information indicating the actual sales value of store A on date DA11. Further, the data DT1-1 includes store history information LG11-0, store history information LG12-0 of store B, weather information WDT10, and search information SDT10 at a date and time DA10 earlier than the date and time DA11 corresponding to the correct information RDT1-1. Include as historical information. In this way, the selection device 100 generates a model using store history information, weather information, and search information at a date and time before the date and time corresponding to the correct information RDT1-1 as history information. As a result, the selection device 100 can generate a model for predicting future demand by inputting the current store KPI status and search status. As for the weather information, the weather forecast for the same date and time as the correct information may be used. For example, when the correct information is information indicating the actual sales value of store A on date DA11, the selection device 100 may use the weather information on date DA11 as an input. In this case, the selection device 100 may use, as an input, forecasted weather information for the date and time to be forecasted based on information such as weather forecasts.

Note that the learning data storage unit 121 may store various types of information, not limited to the above, depending on the purpose. For example, the learning data storage unit 121 may store information such as user number information indicating the number of users located in area A where A store is located and purchase information indicating user purchases. For example, the learning data storage unit 121 may store information regarding the date and time when the learning data was added. Further, for example, the learning data storage unit 121 may store information indicating by what kind of determination processing each learning data was added. For example, the learning data storage unit 121 may store information indicating whether or not each learning data has been determined by the administrator's selection.

(Model information storage unit 122)
The model information storage unit 122 according to the embodiment stores information about models. For example, the model information storage unit 122 stores model information (model data) generated by the generation process. 6 is a diagram illustrating an example of a model information storage unit according to the embodiment; FIG. The model information storage unit 122 shown in FIG. 6 includes items such as "model ID", "usage", and "model data". Although only the model M1 is illustrated in FIG. 6, a large number of model information such as M2, M3, M4, and M5 may be stored according to each application (prediction target).

Each model may also be assumed to be used as a program module that is part of artificial intelligence software. For example, the model M1 may be assumed to be used as a program module for the application "sales forecast (store A)". For example, the model M1 may be assumed to be used as a program module for predicting the sales of store A. For example, the model data MDT1 may be data used as a program module for use "sales forecast (store A)". For example, model M1 causes a computer to output a quantified value of predicted sales corresponding to a given point in time for store A, based on historical information relating to each of a plurality of stores prior to the given point in time. .

"Model ID" indicates identification information for identifying a model. For example, the model identified by model ID "M1" corresponds to model M1 shown in the example of FIG. "Use" indicates the use of the corresponding model. Also, "model data" indicates the data of the associated corresponding model. For example, "model data" includes information including nodes in each layer, functions adopted by each node, connection relations of nodes, and connection coefficients set for connections between nodes.

For example, in the example shown in FIG. 6, the model (model M1) identified by the model ID "M1" is used for "sales forecast (A store)" and is used for sales forecast of A store. It also indicates that the model data of the model M1 is the model data MDT1.

The model M1 (model data MDT1) includes an input layer in which history information related to each of a plurality of stores before a predetermined time is input, an output layer, and any layer from the input layer to the output layer. including a first element belonging to a layer other than the output layer and a second element whose value is calculated based on the weight of the first element and the first element; Each element belonging to each layer other than the layer is treated as the first element, and calculation is performed based on the first element and the weight of the first element to calculate the score used to predict the sales (index) at a predetermined point in time for store A. It is a model for making a computer function so that values are output from the output layer.

Here, it is assumed that the model M1 and the like are realized by a regression model represented by "y=a1*x1+a2*x2+ . . . +ai*xi". In this case, for example, the first element included in model M1 corresponds to input data (xi) such as x1 and x2. Also, the weight of the first element corresponds to the coefficient ai corresponding to xi. Here, the regression model can be viewed as a simple perceptron with an input layer and an output layer. When each model is regarded as a simple perceptron, the first element can be regarded as a node of the input layer, and the second element can be regarded as a node of the output layer.

Also assume that the model M1 and the like are realized by a neural network such as a DNN having one or more intermediate layers. In this case, for example, the first element included in the model M1 corresponds to any node of the input layer or the intermediate layer. Also, the second element corresponds to the next node, which is a node to which the value is transmitted from the node corresponding to the first element. Also, the weight of the first element corresponds to the connection coefficient, which is the weight considered for the value transmitted from the node corresponding to the first element to the node corresponding to the second element.

It should be noted that the model information storage unit 122 may store various types of model information according to purposes, not limited to the above.

(Store information storage unit 123)
The store information storage unit 123 according to the embodiment stores various types of information about stores (enterprises). 7 is a diagram illustrating an example of a store information storage unit according to the embodiment; FIG. For example, the shop information storage unit 123 stores information about various shops. The store information storage unit 123 shown in FIG. 7 includes items such as "store ID", "store name", "category", and "location".

"Store ID" indicates identification information for identifying a store. "Store name" indicates a specific store name of the store. For example, an abstract name such as "A store" is shown, but the "store name" is a specific name such as a combination of a store name such as "○○ cafe X town store" and information indicating the location of the store. may be stored. "Category" indicates the category to which the store belongs. "Location" indicates the place where the store is located. The "location" is illustrated as information indicating an abstract code or range such as "area A", but information indicating latitude and longitude or information indicating an address such as "Z town, Y city, X prefecture". There may be.

For example, in the example of FIG. 7, the store identified by the store ID "SH11" has the store name "A store" and belongs to the category "restaurant". Also, for example, in the example of FIG. 7, the A store (store SH11) is located within the area A.

Note that the store information storage unit 123 may store various types of information, not limited to the above, depending on the purpose. For example, when a company has a plurality of stores, the store information storage unit 123 may store each store in association with each other. For example, when a business has stores in multiple areas, the store information storage unit 123 may store the store in association with each area of the business. For example, the selection device 100 may use store information associated with one area in the store information storage unit 123 when predicting an index for one area for one business operator. For example, the selection device 100 may generate a model using historical information about shops associated with one area. For example, the selection device 100 may predict the index using historical information about shops associated with one area. For example, the store information storage unit 123 may store information indicating whether or not each store is permitted to use its own store's information (provide information) for predicting the indicators of other stores. For example, the store information storage unit 123 may store information indicating whether or not store A is permitted to use the information of store A to predict the indices of other stores such as store B and store C. In this case, the selection device 100 may perform selection processing only for stores that have permitted provision of information to other stores (permitted stores). For example, the selection device 100 uses the information of the one permitted store and the information of the other permitted stores to calculate the predicted value of the index of the one permitted store for the one permitted store that is permitted to provide information to the other stores. may be predicted and selection processing may be performed. As a result, the selection device 100 encourages each store to provide information, enables information collection from a large number of authorized stores, and uses the information to improve prediction accuracy. Therefore, the selection device 100 can improve the quality of services provided.

(Index information storage unit 124)
The index information storage unit 124 according to the embodiment stores various types of information regarding indices. 8 is a diagram illustrating an example of an index information storage unit according to the embodiment; FIG. For example, the index information storage unit 124 stores information about indices that are KPIs of the store. The index information storage unit 124 shown in FIG. 8 has items such as "store", "index", "date and time", "planned value", "measured value", and "predicted value".

"Store" indicates a target store. "Index" indicates an index that is the object of project planning, information collection, and prediction. "Date and time" indicates the time (date and time) for which information is to be collected and predicted. The “planned value” indicates the planned value of the indicator scheduled (planned) for the corresponding day. “Actual value” indicates the actual measured value of the indicator that was actually collected (measured) on the corresponding day. "Predicted value" indicates the predicted value of the predicted indicator for the corresponding day.

For example, in the example shown in FIG. 8, the index is "sales" for store A (store SH11). For the date DA10, the planned sales value is the planned sales value CV10, and the measured sales value is the measured sales value MV10. Further, for the date DA10, the predicted value is "-", indicating that the predicted value of the sales of the A store at the date DA10 is not predicted. For date DA12, the planned sales value is planned sales value CV12, the measured sales value is measured sales value MV12, and the predicted sales value is predicted sales value PV12.

It should be noted that the index information storage unit 124 may store various types of information, not limited to the above, depending on the purpose. Although the example of FIG. 8 shows the case where the information on the index is stored for each store, the information on the store may be stored for each index.

(Coping action information storage unit 125)
The coping behavior information storage unit 125 according to the embodiment stores various types of information about coping behaviors. 9 is a diagram illustrating an example of a coping action information storage unit according to the embodiment; FIG. For example, the coping action information storage unit 125 stores various information such as information indicating the target action and conditions for determining the target action. The coping action information storage unit 125 shown in FIG. 9 has items such as "target", "index", "divergence", "coping action ID", and "content".

"Target" indicates the category of target store. "Index" indicates an index used for determining whether to select a coping action. “Difference” indicates a condition (condition value) that serves as a criterion for determining whether or not to select the target action. "Coping action ID" indicates information for identifying a coping action. "Content" indicates the specific content of the coping action identified by the corresponding coping action.

For example, in the example shown in FIG. 9, the category of the target store is "restaurant" and the index is "sales". The conditions (condition values) for judging divergence are condition values SP1 to SP3. It also indicates that the condition value SP2 is greater than the condition value SP1, and the condition value SP3 is greater than the condition value SP2.

It indicates that the target action (target action AT3) identified by the target action ID "AT3" is the coping action when the condition value SP3 is satisfied, that is, when the divergence is equal to or greater than the condition value SP3. Also, the target action AT3 indicates "X% discount" (X is a predetermined numerical value such as 10 or 30), that is, the implementation of a sale such as an X% discount on the price of the product sold at the store.

It also indicates that the coping action when the condition value SP2 is satisfied, that is, when the divergence is equal to or greater than the condition value SP2, is the target action (target action AT2) identified by the target action ID "AT2". Also, the target action AT2 indicates coupon distribution. Specifically, the target action AT2 indicates distribution of coupons to users located within the area corresponding to the store. It also indicates that the target action (target action AT1) identified by the target action ID "AT1" is the coping action when the condition value SP1 is satisfied, that is, when the divergence is equal to or greater than the condition value SP1. Also, the target behavior AT1 indicates advertisement distribution related to a store.

Note that the coping action information storage unit 125 may store various types of information, not limited to the above, depending on the purpose. Although the example of FIG. 9 shows the case where the information about the index is stored for each store, the store information may be stored for each index.

(control unit 130)
Returning to the description of FIG. 4, the control unit 130 is a controller, and is stored in the storage device inside the selection device 100 by, for example, a CPU (Central Processing Unit) or MPU (Micro Processing Unit). Various programs (corresponding to an example of a selection program) are realized by being executed using 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). The control unit 130 has an input layer and an output layer to which history information related to each of the plurality of shops before a predetermined point in time is inputted by information processing according to the model M1 and the like stored in the model information storage unit 122. and a first element belonging to a layer other than the output layer, which is any layer from the input layer to the output layer, and a second element whose value is calculated based on the first element and the weight of the first element. , with respect to the history information input to the input layer, with each element belonging to each layer other than the output layer as the first element, and performing an operation based on the first element and the weight of the first element, A store The output layer outputs score values used to predict sales at a given point in time.

As shown in FIG. 4, the control unit 130 includes an acquisition unit 131, a generation unit 132, a prediction unit 133, a selection unit 134, and a provision 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. The control unit 130 performs information processing according to the model M1 (model data MDT1) stored in the storage unit 120, a second element whose value is calculated based on the first element and the weight of the first element; , and for the history information input to the input layer, each element belonging to each layer other than the output layer is set as the first element, and an operation based on the first element and the weight of the first element is performed. The output layer outputs score values that are used to predict sales at a given point in time.

(Acquisition unit 131)
Acquisition unit 131 acquires various types of information. For example, the acquisition unit 131 acquires various information from the learning data storage unit 121, the model information storage unit 122, the store information storage unit 123, the index information storage unit 124, the coping behavior information storage unit 125, and the like. Further, the acquisition unit 131 may acquire various types of information from an external information processing device. Further, the acquisition unit 131 may acquire various types of information from the store terminal 20 or the like.

Further, the acquisition unit 131 acquires learning data as shown in the learning data storage unit 121 . For example, the acquiring unit 131 acquires learning data including data DT1-1 to DT1 to DT4 from the learning data storage unit 121. FIG.

The acquisition unit 131 acquires a planned value of an index related to the business of one business operator at a predetermined point in time. The acquisition unit 131 acquires history information related to one business operator. The acquisition unit 131 acquires a planned value of an index related to the business of one business operator, which corresponds to a predetermined point in time, and history information related to the one business operator before the predetermined point in time. . Acquisition unit 131 acquires a planned value of an index related to the business of one store, which corresponds to a predetermined point in time, and history information related to each of a plurality of stores before the predetermined point in time. . The acquisition unit 131 acquires a model that outputs a score indicating a predicted value of one business operator at a predetermined point in time in response to input of history information. Acquisition unit 131 acquires a model that outputs a score indicating a predicted value of one store at a predetermined point in time in response to input of history information.

The acquisition unit 131 acquires history information of a plurality of businesses including one business. Acquisition unit 131 acquires history information of a plurality of stores including one store. The acquisition unit 131 acquires history information including a user's action history before a predetermined point in time. The acquisition unit 131 acquires history information including a user's action history on the Internet. Acquisition unit 131 acquires history information including a search history by a user. Acquisition unit 131 acquires history information including the purchase history of the user. The acquisition unit 131 acquires history information including information on the number of users in an area corresponding to one business operator based on user location information. The acquisition unit 131 acquires history information including information about the number of users in an area corresponding to one store based on user location information. The acquisition unit 131 acquires history information including weather information of an area corresponding to one business operator. Acquisition unit 131 acquires history information including weather information of an area corresponding to one store.

For example, the acquisition unit 131 acquires sales plan information including sales plan values CV10 to CV16 corresponding to dates DA10 to DA16, as shown in the index information storage unit 124 (see FIG. 8). For example, the acquisition unit 131 acquires sales plan information from the store terminal 20 (see FIG. 3) of store A, and stores it in the index information storage unit 124 (see FIG. 8).

Acquisition unit 131 acquires history information of each store from a plurality of stores. The acquisition unit 131 acquires history information related to store KPIs at each date and time (period). For example, the acquisition unit 131 acquires history information regarding the sales, the number of visitors, the number of inquiries, etc. of store A at each date and time (period).

In the example of FIG. 1, the acquisition unit 131 acquires history information of A store from A store. The acquisition unit 131 acquires history information of the A store from the store terminal 20 of the A store. For example, the acquiring unit 131 acquires store history information LG11 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 1, the acquisition unit 131 acquires store history information LG11-0 to LG11-3 corresponding to dates and times DA10 to DA13 from store A. FIG.

In addition, the acquisition unit 131 acquires the history information of the B store from the B store. The acquisition unit 131 acquires the history information of the B store from the store terminal 20 of the B store. For example, the acquiring unit 131 acquires store history information LG12 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 1, the acquiring unit 131 acquires store history information LG12-0 to LG12-3 corresponding to dates and times DA10 to DA13 from store B. FIG.

The acquisition unit 131 also acquires correct information from the A store. The acquisition unit 131 acquires the correct answer information from the store terminal 20 of the A store. For example, the acquisition unit 131 acquires the correct information RDT1 indicating the measured value corresponding to each date and time from the A store. In the example of FIG. 1, the acquiring unit 131 acquires the correct information RDT1-1 to RDT1-4 corresponding to the dates and times DA11 to DA14 from the A store.

The acquisition unit 131 also acquires history information including weather information and information on user behavior (hereinafter also referred to as “user-related information”) from the service providing apparatus 50 . For example, the acquisition unit 131 acquires the history information LG21 corresponding to each date and time from the service providing device 50 . In the example of FIG. 1, the acquisition unit 131 acquires the history information LG21 including weather information and user-related information corresponding to dates DA10 to DA13 from the service providing device 50. In the example of FIG.

For example, the acquisition unit 131 acquires weather information related to the location of store A as weather information. For example, the acquisition unit 131 acquires weather information of area A corresponding to each date and time as weather information related to the location of store A.

For example, the acquisition unit 131 acquires information such as search information indicating a user's search and purchase information indicating a user's purchase as user-related information. The purchase information may be purchase information related to products sold by store A or products related to the products. In addition, the acquisition unit 131 acquires user number information indicating the number of users located in area A where A store is located, as user-related information.

In the example of FIG. 2, the acquisition unit 131 acquires history information of each store from a plurality of stores. The acquisition unit 131 acquires history information related to store KPIs at date DA15. For example, the acquisition unit 131 acquires history information related to the sales, the number of visitors, the number of inquiries, etc. of store A on date and time DA15.

The acquisition unit 131 acquires history information of the A store from the A store. The acquisition unit 131 acquires history information of the A store from the store terminal 20 of the A store. For example, the acquiring unit 131 acquires store history information LG11 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 2, the acquiring unit 131 acquires the store history information LG11-5 corresponding to the date and time DA15 from the A store.

In addition, the acquisition unit 131 acquires the history information of the B store from the B store. The acquisition unit 131 acquires the history information of the B store from the store terminal 20 of the B store. For example, the acquiring unit 131 acquires store history information LG12 including the number of visitors corresponding to each date and time from the A store. In the example of FIG. 2, the obtaining unit 131 obtains the store history information LG12-5 corresponding to the date and time DA15 from the B store.

The acquisition unit 131 also acquires history information including weather information and user-related information from the service providing device 50 . For example, the acquisition unit 131 acquires history information LG25 including weather information WDT15 and search information SDT15 corresponding to date and time DA15 from the service providing device 50 . For example, the acquisition unit 131 acquires weather information related to the location of store A as weather information. For example, the acquisition unit 131 acquires weather information of area A corresponding to each date and time as weather information related to the location of store A. Obtain weather information WDT15 and search information SDT15.

For example, the acquisition unit 131 acquires information such as search information indicating a user's search and purchase information indicating a user's purchase as user-related information. In addition, the acquisition unit 131 acquires user number information indicating the number of users located in area A where A store is located, as user-related information. In addition, as shown in the plan list CL11, the acquisition unit 131 acquires information indicating that the planned sales value of store A on date DA16 is the planned sales value CV16 from the index information storage unit 124 (see FIG. 8). do.

(Generating unit 132)
The generator 132 generates various types of information. For example, the generation unit 132 generates a model as shown in the model information storage unit 122 using learning data stored in the learning data storage unit 121 . For example, the generation unit 132 generates a model that is used for predicting demand related to a business target of a predetermined business operator based on various information acquired by the acquisition unit 131 . For example, based on the learning data acquired by the acquisition unit 131, the generation unit 132 generates a model that is used to predict the demand for the business target of a predetermined business operator. For example, the generation unit 132 generates one Generate a model that is used to predict the sales (index) corresponding to a given point in time for the store.

For example, the generation unit 132 generates the model M1 and the like, and stores the generated model M1 and the like in the model information storage unit 122 . Note that the generation unit 132 may generate the model M1 or the like using any learning algorithm. For example, the generator 132 generates the model M1 and the like using learning algorithms such as neural networks, support vector machines (SVM), clustering, and reinforcement learning. As an example, when the generation unit 132 generates the model M1 or the like using a neural network, the model M1 or the like includes an input layer including one or more neurons, an intermediate layer including one or more neurons, and one or more neurons. and an output layer containing

The generation unit 132 generates a model and stores the generated model in the model information storage unit 122 . Specifically, the generation unit 132 includes an input layer in which history information related to each of a plurality of shops before a predetermined time is input, an output layer, and any layer from the input layer to the output layer. a first element belonging to a layer other than the output layer, and a second element whose value is calculated based on the first element and the weight of the first element, and for the history information input to the input layer, Each element belonging to each layer other than the output layer is treated as the first element, and a calculation based on the first element and the weight of the first element is performed to obtain the score value used to predict the sales at a predetermined point in time for store A. Generate a model that outputs from the output layer.

In the example of FIG. 1, the generation unit 132 generates learning data using the actual sales values (measured sales values) for each date and time DA11 to DA14 included in the correct information RDT1 as the correct information RDT1-1 to RDT1-4. . For example, the generation unit 132 adds learning data to the learning data storage unit 121 .

The generation unit 132 generates, as learning data, data DT1-1 including correct information RDT1-1 indicating the actual sales value of store A on date and time DA11. The generation unit 132 also generates data DT1-1 including store history information of store A and store B prior to date and time DA11 as learning data. In the example of FIG. 1, the generation unit 132 generates data DT1-1 including store history information LG11-0 of store A and store history information LG12-0 of store B corresponding to date and time DA10 as learning data. The generation unit 132 also generates data DT1-1 including weather information and user-related information prior to the date and time DA11 as learning data. In the example of FIG. 1, the generation unit 132 generates data DT1-1 including, as learning data, weather information WDT10 and search information SDT10 at date DA10 earlier than date DA11, among history information LG21.

Also, for example, the generation unit 132 generates data DT1-2 including correct information RDT1-2 indicating the actual sales value of store A on date and time DA12 as learning data. The generation unit 132 also generates data DT1-2 including store history information of store A and store B prior to date and time DA12 as learning data. In the example of FIG. 1, the generation unit 132 generates data DT1-2 including store history information LG11-1 of store A and store history information LG12-1 of store B corresponding to date and time DA11 as learning data. Further, the generation unit 132 generates data DT1-2 including weather information and user-related information prior to date and time DA12 as learning data. In the example of FIG. 1, the generation unit 132 generates data DT1-2 including, as learning data, the weather information WDT11 and the search information SDT11 at the date and time DA11 earlier than the date and time DA12 from the history information LG21.

In the example of FIG. 1, the generator 132 generates a model based on learning data. For example, the generator 132 generates a model based on learning data. For example, the generation unit 132 generates a model by performing learning using the data DT1-1 to DT1-4 in the learning data storage unit 121 as learning data (teaching data). The generation unit 132 generates a model using store history information, weather information, and user-related information at a date and time before the date and time corresponding to each correct answer information as history information.

The generation unit 132 generates the model M1 using learning data as shown in the learning data storage unit 121 . For example, when the correct information RDT1-1 is "1" indicating that the sales were 100,000 yen or more, the generating unit 132 generates , the learning process is performed so that the score output by the model M1 approaches "1". For example, when store history information LG11-0, store history information LG12-0, weather information WDT10, and search information SDT10 of data DT1-1 are input to model M1, generating unit 132 determines that the score output by model M1 is Learning processing is performed so as to approach "1".

Further, when the correct information RDT1-2 is "0.5" indicating that the sales were 50,000 yen, the generating unit 132 generates the model M1 when the history information included in the data DT1-2 is input. Then, the learning process is performed so that the score output by the model M1 approaches "0.5". For example, when store history information LG11-1, store history information LG12-1, weather information WDT11, and search information SDT11 of data DT1-2 are input to model M1, generating unit 132 determines that the score output by model M1 is Learning processing is performed so as to approach "0.5".

The generation unit 132 generates input data by combining information acquired by the acquisition unit 131 . In the example of FIG. 2, the generation unit 132 inputs data DT21 including store history information LG11-5 of store A and store history information LG12-5 of store B corresponding to date and time DA15, as shown in input data list IPD21. Generate as data. The generation unit 132 also generates data DT21 including weather information and user-related information prior to date and time DA16 as input data. As shown in the input data list IPD21, the generation unit 132 generates data DT1-1 including weather information WDT15 and search information SDT15 as history information as learning data.

(Prediction unit 133)
The prediction unit 133 predicts various information. The prediction unit 133 uses the information stored in the learning data storage unit 121, the model information storage unit 122, the shop information storage unit 123, the index information storage unit 124, the coping behavior information storage unit 125, and the like to perform various Anticipate information. For example, the prediction unit 133 predicts various information based on various information acquired by the acquisition unit 131 .

Based on the history information acquired by the acquisition unit 131, the prediction unit 133 predicts a prediction value of an index related to the business of one business operator at a predetermined point in time. Based on the history information acquired by the acquisition unit 131, the prediction unit 133 predicts a predicted value of an index related to the business of one store at a predetermined point in time. The prediction unit 133 inputs history information into the model, and predicts a prediction value of one business operator at a predetermined point in time based on the score output by the model. The prediction unit 133 inputs the history information into the model and predicts the predicted value of one store at a predetermined point in time based on the score output by the model.

The prediction unit 133 predicts a predicted value of one store at a predetermined point in time using history information of a plurality of stores including one store. The prediction unit 133 predicts the predicted value of one store at a predetermined time using history information including the user's action history before the predetermined time. The prediction unit 133 predicts the predicted value of one store at a predetermined point in time using history information including the user's action history on the Internet. The prediction unit 133 predicts a predicted value of one store at a predetermined point in time using history information including a user's search history. The prediction unit 133 predicts the predicted value of one store at a predetermined point in time using history information including the purchase history of the user. The prediction unit 133 predicts a predicted value of one store at a predetermined point in time using history information including information about the number of users in an area corresponding to one store based on user position information. The prediction unit 133 predicts a predicted value for one store at a predetermined point in time using history information including weather information for an area corresponding to one store.

In the example of FIG. 2, the prediction unit 133 inputs data DT21 into the model. The prediction unit 133 calculates the score indicating the sales of the A store on the date and time DA16 by performing the processing shown in the processing group PS21. The prediction unit 133 inputs the data DT21 to the model M1. The prediction unit 133 inputs history information including store history information LG11-5, LG12-5, weather information WDT15, search information SDT15, etc. to the model M1. Model M1 to which data DT21 is input outputs a score. The prediction unit 133 predicts the predicted sales value for date DA16 based on the score indicating the sales of store A on date DA16. The prediction unit 133 predicts the predicted sales value for date/time DA16 using a correspondence table showing the correspondence between each level and sales. In the example of FIG. 2, as shown in the prediction list PL11, a correspondence table showing the correspondence between each level and sales is used to predict the sales forecast value PV16 for the date DA16. For example, when the score SC16 is "0.2", the selection device 100 uses the correspondence table showing the correspondence between each level and the sales, and the sales of the A store predicted on the date and time DA16 is 20,000 yen (for example, The sales forecast value PV16) is predicted.

For example, the prediction unit 133 calculates the score using a model having an arbitrary structure such as the regression model or neural network described above. Specifically, the model M1 predicts a predetermined target when history information related to each of a plurality of stores before a predetermined time (that is, each element used for calculating the score described above) is input. is set so as to output a quantified value (that is, a score suggesting how much sales are at store A at a predetermined point in time). The prediction unit 133 uses such a model M1 to calculate a score related to the predicted sales value of store A at a predetermined point in time.

In the above example, when history information related to each of a plurality of stores before a predetermined point in time is input, the model M1 quantifies the predicted sales of store A at a given point in time (predicted value ) is shown as an example. However, the model (model X) according to the embodiment may be a model generated based on results obtained by repeatedly inputting and outputting data to the model M1. For example, the model X may be a model (model Y) that has been learned to input historical information related to each of a plurality of stores before a predetermined point in time and to output the score output by the model M1. Alternatively, the model M1 may be a model that has been learned such that historical information related to each of a plurality of stores before a predetermined time is input and the output value of the model Y is output. Further, when the prediction unit 133 performs prediction processing using a GAN (Generative Adversarial Networks), the model M1 may be a model forming part of the GAN.

(Selection unit 134)
The selection unit 134 selects various types of information. The selection unit 134 uses the information stored in the learning data storage unit 121, the model information storage unit 122, the shop information storage unit 123, the index information storage unit 124, the coping behavior information storage unit 125, and the like to perform various Select information. For example, the selection unit 134 selects various information from information stored in the learning data storage unit 121, the model information storage unit 122, the store information storage unit 123, the index information storage unit 124, the coping behavior information storage unit 125, and the like. Select information for

The selection unit 134 selects a coping action, which is an action for coping with the deviation, when the deviation between the predicted value predicted by the prediction unit 133 and the planned value satisfies a predetermined condition. The selection unit 134 compares the planned value and the predicted value to determine whether the divergence satisfies a predetermined condition, and selects a coping action when it is determined that the predetermined condition is satisfied. The selection unit 134 selects a coping action when the difference between the planned value and the predicted value is greater than or equal to a predetermined threshold. The selection unit 134 selects one coping action from among a plurality of coping actions related to improvement of the index of one business operator according to the degree of divergence between the planned value and the predicted value. The selection unit 134 selects one coping action from among a plurality of coping actions relating to improvement of the index of one store according to the degree of divergence between the planned value and the predicted value.

In the example of FIG. 2, the selection unit 134 calculates the difference between the planned sales value and the predicted sales value for the date and time DA16. The selection unit 134 calculates that the difference between the planned sales value CV16 and the predicted sales value PV16 is the difference DF11, as shown in the deviation list DL11. For example, the selection unit 134 calculates the difference DF11 by subtracting the predicted sales value PV16 from the planned sales value CV16.

The selection unit 134 determines whether the difference DF11 indicating the deviation between the planned sales value CV16 and the predicted sales value PV16 satisfies the condition. For example, the selection unit 134 uses the conditions corresponding to the category “eating and drinking” and the index “sales” of store A among the information stored in the coping action storage unit 125 in FIG. 1, and the difference DF11 satisfies the conditions. determine whether For example, the selection unit 134 determines which of the deviations between the object “eating and drinking” and the index “sales” stored in the coping action storage unit 125 shown in FIG. 1 corresponds to which deviation condition.

The selection unit 134 selects one coping action from a plurality of coping actions according to the degree of divergence between the planned value and the predicted value. The selection unit 134 selects, as a coping action, one measure from among a plurality of measures relating to improvement of the index of one store according to the degree of divergence between the planned value and the predicted value. For example, the selection unit 134 selects one of the coping actions AT1 to AT3 as the coping action according to the magnitude of the difference between the planned value and the predicted value. For example, when the difference between the planned value and the predicted value is equal to or greater than the condition value SP3, the selection unit 134 selects the X% discount (coping action AT3), which is one of the coping actions AT1 to AT3. For example, when the difference between the planned value and the predicted value is smaller than the condition value SP3 and equal to or greater than the condition value SP2, the selection unit 134 selects coupon distribution (coping action AT2) as one measure among the coping actions AT1 to AT3. select. For example, when the difference between the planned value and the predicted value is smaller than the condition value SP2 and equal to or greater than the condition value SP1, the selection unit 134 selects advertisement delivery (coping action AT1) as one measure among the coping actions AT1 to AT3. select.

In the example of FIG. 2, the selection unit 134 determines that the difference DF11 is greater than the condition value SP2 and less than the condition value SP3 by comparing the difference DF11 with the condition values SP1 to SP3. In this case, the selection unit 134 determines that the divergence between the planned sales value and the predicted sales value of the A store satisfies the divergence condition of the condition value SP2. Then, the selection unit 134 selects a coping action based on the deviation between the planned value and the predicted value. In the example of FIG. 2, the selection unit 134 selects the coping action AT2 corresponding to the condition value SP2 because the difference between the planned sales value and the predicted sales value of the store A satisfies the conditions for the difference of the condition value SP2.

(Providing unit 135)
The providing unit 135 provides various services. For example, the providing unit 135 provides various information. For example, the providing unit 135 transmits various information to the terminal device 10 of the user. For example, the providing unit 135 transmits various information to the store terminal 20 . Also, for example, the providing unit 135 transmits various types of information to the service providing apparatus 50 .

For example, the providing unit 135 provides various services based on the coping action selected by the selecting unit 134 . For example, when the coping action selected by the selection unit 134 is to provide information to the terminal device 10 of the user, the provision unit 135 provides an information provision service to the terminal device 10 . For example, the provision unit 135 distributes (transmits) an advertisement to the terminal device 10 when the coping action selected by the selection unit 134 is to distribute an advertisement to the terminal device 10 of the user. For example, when the coping action selected by the selection unit 134 is to distribute a coupon such as a discount to the terminal device 10 of the user, the providing unit 135 distributes (transmits) a coupon such as a discount to the terminal device 10 . Further, for example, when the coping action selected by the selection unit 134 is to provide information to the shop terminal 20 of the shop, the providing unit 135 provides an information providing service to the shop terminal 20 . For example, when the coping action selected by the selection unit 134 is execution of a sale such as a discount, the provision unit 135 distributes (transmits) information requesting execution of the sale to the store terminal 20 .

Further, when the service providing device 50 provides information to the terminal device 10 or the store terminal 20 , the providing unit 135 provides the service providing device 50 with information indicating the coping action selected by the selecting unit 134 . For example, when the coping action selected by the selection unit 134 is to provide information to the terminal device 10 of the user, the provision unit 135 requests the service provision device 50 to provide information provision service to the terminal device 10 . For example, when the coping action selected by the selection unit 134 is to distribute an advertisement to the terminal device 10 of the user, the provision unit 135 transmits information requesting distribution (transmission) of the advertisement to the terminal device 10 as a service provision. Send to device 50 . For example, when the coping action selected by the selection unit 134 is to distribute a coupon such as a discount to the terminal device 10 of the user, the providing unit 135 requests distribution (transmission) of a coupon such as a discount to the terminal device 10. information to be sent to the service providing apparatus 50 . Further, for example, when the coping action selected by the selection unit 134 is to provide information to the shop terminal 20 of the shop, the providing unit 135 requests the service providing device 50 to provide the information providing service to the shop terminal 20. do. For example, when the coping action selected by the selection unit 134 is execution of a sale such as a discount, the provision unit 135 requests the store terminal 20 to distribute (transmit) information requesting execution of the sale. Information is transmitted to the service providing device 50 .

In the example of FIG. 2 , the provision unit 135 executes the coping action selected by the selection unit 134 . In the example of FIG. 2, the providing unit 135 requests the service providing apparatus 50 to distribute coupons to users located in area A where store A is located. The providing unit 135 transmits to the service providing apparatus 50 request information requesting the service providing apparatus 50 to distribute coupons to users located in the area A where the store A is located.

[4. Generation process flow]
Next, a procedure of generation processing by the selection system 1 according to the embodiment will be described with reference to FIG. 10 . FIG. 10 is a flowchart illustrating an example of generation processing according to the embodiment.

As shown in FIG. 10, the selection device 100 acquires learning data (step S101). For example, the selection device 100 acquires learning data from the learning data storage unit 121 .

After that, the selection device 100 generates a model based on the learning data (step S102). In the example of FIG. 1, the selection device 100 uses learning data from the learning data storage unit 121 to generate the model M1.

[5. Selection process flow]
Next, a procedure of generation processing by the selection system 1 according to the embodiment will be described with reference to FIG. 11 . FIG. 11 is a flowchart illustrating an example of selection processing according to the embodiment.

As shown in FIG. 11, the selection device 100 acquires history information (step S201). In the example of FIG. 2, the selection device 100 acquires the history information LG11-5 at the date and time DA15 from the A store. Further, in the example of FIG. 2, the selection device 100 acquires store history information LG12-5 at the date and time DA15 from the B store. In addition, in the example of FIG. 2, the selection device 100 acquires the history information LG25 at the date and time DA15 from the service providing device 50 .

The selection device 100 also predicts the predicted value of the index using the history information and the model (step S202). The selection device 100 predicts the predicted sales value as an index for the A store using the history information and the model M1. In the example of FIG. 2, the selection device 100 uses data DT21 including history information LG11-5, LG12-5, LG25, etc. on date DA15 and the model M1 to calculate the predicted sales of store A on date DA16. A predicted value PV16 is predicted.

The selection device 100 also acquires the planned value of the index (step S203). In the example of FIG. 2, the selection device 100 acquires information indicating that the planned sales value for store A on date and time DA16 is the planned sales value CV16. For example, the selection device 100 acquires information indicating that the planned sales value of store A on date DA16 is the planned sales value CV16 from the index information storage unit 124 (see FIG. 8).

Note that step S203 may be performed at any stage before step S204. For example, step S203 may be performed before or at the same timing as step S202.

Then, the selection device 100 determines whether the deviation between the planned value and the predicted value satisfies a predetermined condition (step S204). In the example of FIG. 2, the selection device 100 calculates the difference between the planned sales value CV16 and the predicted sales value PV16. In this case, the selection device 100 calculates that the difference between the planned sales value CV16 and the predicted sales value PV16 is the difference DF11. Then, the selection device 100 determines to which condition the difference DF11 indicating the deviation between the planned sales value CV16 and the predicted sales value PV16 corresponds. For example, the selection device 100 determines which of the divergence conditions between the object “eating and drinking” and the index “sales” stored in the coping action storage unit 125 (see FIG. 9) corresponds to which divergence condition.

In the example of FIG. 2, the selection device 100 determines that the difference DF11 is greater than the condition value SP2 and less than the condition value SP3 by comparing the difference DF11 with the condition values SP1 to SP3. In this case, the selection device 100 determines that the divergence between the planned sales value and the predicted sales value of the store A satisfies the divergence condition of the condition value SP2.

When determining that the deviation between the planned value and the predicted value satisfies the predetermined condition (step S204: Yes), the selection device 100 selects a coping action based on the deviation between the planned value and the predicted value (step S205). ). In the example of FIG. 2, the selection device 100 selects the coping action AT2 corresponding to the condition value SP2 because the difference between the planned sales value and the predicted sales value of the store A satisfies the conditions for the difference of the condition value SP2.

The selection device 100 executes the selected coping action (step S206). For example, when the coping action AT2 is to distribute coupons to users within the area, the selection device 100 executes coupon distribution to users located in area A where store A is located. In the example of FIG. 2, the selection device 100 requests the service providing device 50 to distribute coupons to users located in area A where A store is located.

On the other hand, when the selection device 100 determines that the divergence between the planned value and the predicted value does not satisfy the predetermined condition (step S204: No), the process ends. For example, when the difference between the planned sales value and the predicted sales value of store A is less than the condition value SP1, the selection device 100 determines that none of the conditions are met, and ends the process without selecting any coping action.

[6. effect〕
As described above, the selection device 100 according to the embodiment has the acquisition unit 131, the prediction unit 133, and the selection unit . The acquisition unit 131 acquires a planned value of an index related to the business of one business operator, which corresponds to a predetermined point in time, and history information related to the one business operator before the predetermined point in time. . Based on the history information acquired by the acquisition unit 131, the prediction unit 133 predicts a prediction value of an index related to the business of one business operator at a predetermined point in time. The selection unit 134 selects a coping action, which is an action for coping with the deviation, when the deviation between the predicted value predicted by the prediction unit 133 and the planned value satisfies a predetermined condition.

As described above, when the deviation between the predicted value and the planned value satisfies a predetermined condition, the selection device 100 according to the embodiment selects a coping action, which is an action for coping with the deviation, thereby performing a predetermined business. It can enable appropriate actions about a person.

Also, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information related to one business operator.

In this way, the selection device 100 according to the embodiment uses historical information related to one business operator to predict a predicted value of one business operator at a predetermined point in time, can enable appropriate actions for a given operator.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires a model that outputs a score indicating a predicted value of one business operator at a predetermined point in time in response to input of history information. The prediction unit 133 inputs history information into the model, and predicts a prediction value of one business operator at a predetermined point in time based on the score output by the model.

In this way, the selection device 100 according to the embodiment uses the model to predict the predicted value of one business operator at a predetermined point in time based on the score output by the model, and responds based on the predicted predicted value. Selecting an action may enable the appropriate action for a given operator.

In addition, in the selection device 100 according to the embodiment, the selection unit 134 determines whether or not the divergence satisfies a predetermined condition by comparing the planned value and the predicted value, and when it is determined that the predetermined condition is satisfied. to select a coping action.

In this way, the selection device 100 according to the embodiment compares the planned value and the predicted value to determine whether or not the deviation satisfies the predetermined condition, and if it determines that the predetermined condition is satisfied, the selection device 100 takes action. Selecting an action may enable the appropriate action for a given operator.

In addition, in the selection device 100 according to the embodiment, the selection unit 134 selects a coping action when the difference between the planned value and the predicted value is equal to or greater than a predetermined threshold.

As described above, the selection device 100 according to the embodiment selects a coping action when the difference between the planned value and the predicted value is equal to or greater than a predetermined threshold value, thereby enabling an appropriate action for a predetermined business operator. be able to.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information of a plurality of businesses including one business.

As described above, the selection device 100 according to the embodiment can improve the prediction accuracy of the predicted value by predicting the predicted value using history information of a plurality of businesses including one business operator. Appropriate actions can be enabled for a given operator.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information including the user's action history before a predetermined point in time.

As described above, the selection device 100 according to the embodiment can improve the prediction accuracy of the predicted value by predicting the predicted value using history information including the user's action history before a predetermined point in time. Appropriate actions can be enabled for a given operator.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information including the user's action history on the Internet.

As described above, the selection device 100 according to the embodiment can improve the prediction accuracy of the prediction value by predicting the prediction value using history information including the behavior history of the user on the Internet. It can enable appropriate actions about a person.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information including search history by the user.

As described above, the selection device 100 according to the embodiment can improve the prediction accuracy of the prediction value by predicting the prediction value using the history information including the search history of the user. It can enable appropriate behavior.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information including the purchase history of the user.

As described above, the selection device 100 according to the embodiment can improve the prediction accuracy of the predicted value by predicting the predicted value using history information including the purchase history of the user. It can enable appropriate behavior.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information including information regarding the number of users in an area corresponding to one store based on user position information.

In this way, the selection device 100 according to the embodiment predicts a predicted value using history information including information about the number of users in an area corresponding to one store based on user location information. Accuracy can be improved, so appropriate action can be enabled for a given operator.

In addition, in the selection device 100 according to the embodiment, the acquisition unit 131 acquires history information including weather information of an area corresponding to one store.

As described above, the selection device 100 according to the embodiment can improve the prediction accuracy of the predicted value by predicting the predicted value using the history information including the weather information of the area corresponding to one store. , can enable appropriate actions for a given operator.

In addition, in the selection device 100 according to the embodiment, the selection unit 134 selects one coping action from a plurality of coping actions according to the degree of divergence between the planned value and the predicted value.

In this way, the selection device 100 according to the embodiment selects one coping action from a plurality of coping actions according to the degree of divergence between the planned value and the predicted value. actions can be enabled.

[7. Hardware configuration]
The selection device 100 according to the embodiments described above is implemented by a computer 1000 configured as shown in FIG. 12, for example. FIG. 12 is a hardware configuration diagram showing an example of a computer that implements the functions of the selection device. 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 to other devices via network N. FIG.

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 . CPU 1100 also 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 selection device 100 according to the embodiment, the CPU 1100 of the computer 1000 executes a program or data (for example, model M1 (model data MDT1)) loaded on the RAM 1200 to control It implements the function of the unit 130 . CPU 1100 of computer 1000 reads these programs or data (for example, model M1 (model data MDT1)) from recording medium 1800 and executes them. may be obtained.

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

[8. 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 acquisition unit can be read as acquisition means or an acquisition circuit.

1 selection system 100 selection device 121 learning data storage unit 122 model information storage unit 123 store information storage unit 124 index information storage unit 125 coping behavior information storage unit 130 control unit 131 acquisition unit 132 generation unit 133 prediction unit 134 selection unit 135 provision unit 10 terminal device 20 store terminal 50 service providing device N network

Claims (10)

A planned value of an index related to the business of one business operator, which is a planned value corresponding to a predetermined time point, is obtained from the business operator device used by the one business operator, and history information before the predetermined time point is obtained. an acquisition unit that acquires a model that is acquired from the business operator device or another external device and that outputs a score indicating the level of the indicator of the one business operator at a predetermined point in time according to the input of history information;
The history information acquired by the acquisition unit is input to the model, and the score output by the model and a correspondence table showing the correspondence between each level of the score and the index are used to determine the business of the one business operator. a prediction unit that predicts a predicted value of the associated indicator, the predicted value corresponding to the predetermined point in time;
The plurality of a selection unit that selects one coping action from among the coping actions;
with
The acquisition unit
A selection device that acquires history information of a plurality of businesses including the history information of the one business. The acquisition unit
2. The selection according to claim 1, wherein the history information related to one business operator is obtained, wherein the historical information includes the number of visitors corresponding to each date and time to the store of the one business operator. Device. The acquisition unit
3. The selection device according to claim 1, wherein the history information including the user's action history before the predetermined time is acquired. The acquisition unit
4. The selection device according to claim 3, wherein said history information including a user's action history on the Internet is obtained. The acquisition unit
5. The selection device according to claim 3, wherein the history information including search history by the user is obtained. The acquisition unit
6. The selection device according to any one of claims 3 to 5, wherein said history information including purchase history by said user is obtained. The acquisition unit
7. The history information including information indicating the number of users located in the area corresponding to the one operator based on the location information of the users is obtained according to any one of claims 3 to 6. selection device. The acquisition unit
8. The selection device according to any one of claims 1 to 7, wherein said history information including weather information of an area corresponding to said one operator is obtained. A computer implemented selection method comprising:
A planned value of an index related to the business of one business operator, which is a planned value corresponding to a predetermined time point, is obtained from the business operator device used by the one business operator, and history information before the predetermined time point is obtained. an acquisition step of acquiring a model that is acquired from the business operator device or another external device and that outputs a score indicating the level of the index of the one business operator at a predetermined point in time according to the input of history information;
The history information acquired in the acquisition step is input to the model, and the score output by the model and a correspondence table showing the correspondence between each level of the score and the index are used to determine the business of the one business operator. a prediction step of predicting a predicted value of the associated indicator, the predicted value corresponding to the predetermined point in time;
The plurality of a selection step of selecting one coping action from among the coping actions;
including
The obtaining step includes
A selection method characterized by acquiring history information of a plurality of businesses including the history information of the one business. A planned value of an index related to the business of one business operator, which is a planned value corresponding to a predetermined time point, is obtained from the business operator device used by the one business operator, and history information before the predetermined time point is obtained. an acquisition procedure for acquiring a model that is acquired from the business operator device or another external device and that outputs a score indicating the level of the index of the one business operator at a predetermined point in time according to input of history information;
The history information acquired by the acquisition procedure is input to the model, and the score output by the model and a correspondence table showing the correspondence between each level of the score and the indicator are used to determine the business of the one business operator. a prediction procedure for predicting a predicted value of the associated indicator, the predicted value corresponding to the given point in time;
The plurality of A selection procedure for selecting one coping action from among the coping actions;
on the computer, and
The acquisition procedure includes:
A selection program for acquiring history information of a plurality of businesses including the history information of the one business.

Images