JP6194092B1 - Calculation device, calculation method, and calculation program - Google Patents

Abstract

[PROBLEMS] To calculate a trustworthiness in accordance with the actual situation. A calculation apparatus according to the present application includes an acquisition unit and a calculation unit. An acquisition part acquires the information regarding the shopkeeper who lost the right to perform a commercial transaction in the predetermined | prescribed commercial transaction site on a network. The calculation unit calculates a credit rating for the store clerk opening the store on a predetermined commercial transaction site based on the information regarding the store clerk acquired by the acquisition unit. For example, an acquisition part acquires the information regarding the evaluation from the user with respect to the store-closing person in a predetermined | prescribed commercial transaction site as information regarding a store-closing person. [Selection] Figure 1

Description

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

  In general, when a financial institution such as a bank decides a loan facility (credit facility), it refers to a business provided by a data provider that specializes in collecting and analyzing business (company) data. The Such business operator data is generated based on, for example, business operators' financial statements (for example, financial statements, profit and loss statements, etc.). A financial institution uses business data provided by a business in order to calculate a credit rating for determining whether or not the loan amount to each business is appropriate.

  Here, not only data with long update intervals (static data) such as financial data published in the fiscal year end, but also models with input of short update intervals (dynamic data) such as stock prices are created. And the system which can perform the newest company evaluation in time is proposed (for example, patent document 1). In addition to the rapid spread of the Internet in recent years, in addition to information based on business operators' financial statements as in the past, information users share information about businesses on the Internet, and the information they share is unified. A technique for enhancing the objectivity of business operator data by performing integrated management is known (for example, Patent Document 2).

JP 2000-348015 A JP 2010-26602 A

  However, in the above-described conventional technology, it is not always possible to calculate the degree of trustworthiness in accordance with the actual situation. Specifically, the above-mentioned conventional technology only shares information such as financial information of publicly disclosed business operators, business results of business operators that information users can know, and industry news. Even if they are integrated, it is difficult to appropriately evaluate the creditworthiness of the operator (for example, the growth potential and stability of the operator over the future). In addition, since data such as stock prices is not disclosed to unlisted companies, in some cases, there is a shortage of data for corporate evaluation, making it difficult to appropriately calculate the creditworthiness of the business operator.

  The present application has been made in view of the above, and an object thereof is to provide a calculation device, a calculation method, and a calculation program capable of calculating a trustworthiness in accordance with the actual situation.

  The calculation device according to the present application is based on an acquisition unit that acquires information on a store clerk who has lost the right to perform a commercial transaction on a predetermined commercial transaction site on a network, and information on the store clerk acquired by the acquisition unit. And a calculation unit for calculating a credit rating for the store clerk opening the predetermined commercial transaction site.

  According to one aspect of the embodiment, there is an effect that the degree of trust according to the actual situation can be calculated.

FIG. 1 is a diagram illustrating an example of a calculation process according to the embodiment. FIG. 2 is a diagram illustrating a configuration example of a calculation system according to the embodiment. FIG. 3 is a diagram illustrating a configuration example of the calculation apparatus according to the embodiment. FIG. 4 is a diagram illustrating an example of a store opening information storage unit according to the embodiment. FIG. 5 is a diagram illustrating an example of a business information table according to the embodiment. FIG. 6 is a diagram illustrating an example of the store information table according to the embodiment. FIG. 7 is a diagram illustrating an example of a product information table according to the embodiment. FIG. 8 is a diagram illustrating an example of a word information table according to the embodiment. FIG. 9 is a diagram illustrating an example of a model storage unit according to the embodiment. FIG. 10 is a diagram illustrating an example of a loan condition storage unit according to the embodiment. FIG. 11 is a flowchart illustrating a generation processing procedure performed by the calculation apparatus according to the embodiment. FIG. 12 is a flowchart illustrating a calculation processing procedure performed by the calculation apparatus according to the embodiment. FIG. 13 is a hardware configuration diagram illustrating an example of a computer that realizes the function of the calculation apparatus.

  Hereinafter, a mode for carrying out a calculation device, a calculation method, and a calculation program according to the present application (hereinafter referred to as “embodiment”) will be described in detail with reference to the drawings. Note that the calculation device, the calculation method, and the calculation program according to the present application are not limited by this embodiment. In the following embodiments, the same portions are denoted by the same reference numerals, and redundant description is omitted.

[1. Example of calculation process)
First, an example of calculation processing according to the embodiment will be described with reference to FIG. FIG. 1 is a diagram illustrating an example of a calculation process according to the embodiment. In FIG. 1, the calculation process by the calculation apparatus 100 according to the embodiment will be described. Specifically, FIG. 1 illustrates an example in which the calculation device 100 performs a calculation process for calculating a credit rating for a store clerk opening a store on a predetermined commercial transaction site on the network.

  In the example illustrated in FIG. 1, the calculation device 100, the user terminal 10, and the web server 30 are communicably connected via a communication network (not shown) (for example, the Internet).

  A calculation apparatus 100 illustrated in FIG. 1 is a server apparatus that performs a calculation process for calculating a credit rating for a business operator. In the embodiment, the calculation device 100 calculates a credit rating for a store operator who opens a shopping mall that is a predetermined commercial transaction site on the network provided by the web server 30. In the embodiment, the calculation device 100 uses the calculated reliability for various processes. For example, the calculation device 100 performs a process of determining a loan condition (credit line, interest rate, etc.) for the business based on the calculated credit quality.

  A user terminal 10 illustrated in FIG. 1 is an information processing apparatus used by a user. Specifically, the user terminal 10 allows a user to browse a web page, purchases a product on a website (for example, a shopping mall on the network), or posts a product information evaluation (so-called review). Used for. In the example of FIG. 1, the user terminal 10 is, for example, a smartphone or a tablet terminal.

  The web server 30 illustrated in FIG. 1 is a server device that provides various web pages when accessed from the user terminal 10. In the example illustrated in FIG. 1, the web server 30 provides a shopping mall site on the network. In a shopping mall provided by the web server 30, each business operator opens a store that handles products of various categories. In the embodiment, a business operator that opens a store in a shopping mall is referred to as a store opener. In addition, a business that has opened a store in a shopping mall and does not open a store in the shopping mall at the present time is referred to as a store exit.

  The web server 30 may acquire information related to user behavior on the network via a website to be provided. Information about user behavior is stored in, for example, a storage unit provided in the web server 30 or a predetermined storage device. Information on user behavior on the network means information transmitted from the user terminal 10 in accordance with a user operation when using a service provided by the web server 30 (a shopping mall in the example of FIG. 1). For example, the user behavior on the network is a purchase behavior in a shopping mall, a posting of a review from a user for each store in the shopping mall, or the like.

  By the way, a business that opens a business transaction site on a network such as a shopping mall provided by the web server 30 has a small business structure, a relatively new establishment, a private listing, or a personality. Often, they are business owners and do not have reliable financial information or experience. In this case, the business operator is determined not to have creditworthiness indicating credit and may not be able to receive a loan from a financial institution such as a bank.

  For example, even if a business operator publishes reliable financial information, general business financial information is published once every quarter or half year, except for information that fluctuates in real time such as the stock price of a listed company. In many cases, For this reason, even if the creditworthiness is calculated based on the financial information of the business operator, the creditworthiness may not reflect the real-time performance on the actual commercial transaction site. In other words, even if there is a possibility that the business will actually be successful in the future, if the future potential is not properly reflected in the creditworthiness, the credit from the financial institution cannot be obtained and the loan can be received. There will be no.

  Therefore, the calculation device 100 according to the embodiment performs a credit calculation process in accordance with the actual situation in a predetermined commercial transaction site such as a shopping mall, so that it is possible for venture companies and small businesses that normally have difficulty in calculating the credit rating. Allows creditworthiness to be calculated. Specifically, the calculation device 100 acquires information about a business operator who has lost the right to perform a commercial transaction in a shopping mall, and calculates a credit rating for a store opening a store in the shopping mall based on the acquired information.

  The business operator who has lost the right to conduct business transactions in the shopping mall refers to a business operator who has opened a store in the shopping mall in the past and then left the shopping mall. In the embodiment, a business operator who has left the shopping mall is referred to as a store exit. For example, a person who has left the store is a business operator who has become unable to operate the business in a shopping mall. Further, for example, the store exit is a business that has been forcibly left by the administrator of the web server 30 due to violation of the store opening rules in the shopping mall. In addition, a business person who has stopped opening a store in a shopping mall for some reason may be included in a store exit.

  In the embodiment, the calculation device 100 calculates the correlation between information regarding a store keeper who has lost the right to perform a commercial transaction in a shopping mall and information regarding a store clerk currently opening in the shopping mall. Then, the calculation device 100 calculates the creditworthiness of the shopkeeper based on the calculated correlation.

  Here, it is presumed that the store clerk who lost the right to conduct a commercial transaction in the shopping mall had some problem that led to the store exit. That is, the calculation apparatus 100 determines the possibility that the store opening person will leave the store in the shopping mall in the future by obtaining the correlation between the information related to the store opening person and the information related to the store opening person. Then, the calculation device 100 calculates a credit rating indicating the credit of the operator based on the determination result. As a result, the calculation device 100 can calculate a trustworthy degree of trust in line with the actual situation even if the store has relatively little track record at the present time or a business operator whose financial information is not disclosed. Hereinafter, the calculation processing according to the embodiment will be described along the flow with reference to FIG. 1.

  In the example illustrated in FIG. 1, it is assumed that a shopping mall provided by the web server 30 includes a web page 61 of a shop AAA operated by a store exit 60 that has opened in the shopping mall in the past. Further, it is assumed that the shopping mall has a web page 71 of a shop BBB operated by a shopkeeper 70 that is currently open. Although not shown in FIG. 1, the shopping mall provided by the web server 30 includes a considerable number of shopkeepers 60 enough to obtain a sufficient amount of information for performing a calculation process described later. It shall exist.

  It is assumed that the user terminal 10 accesses a shopping mall provided by the web server 30 according to an operation by the user. Although not shown in FIG. 1, there are a plurality of user terminals 10, and a sufficient amount is sufficient to perform a calculation process to be described later, from the time when the store 60 has opened to the present time. It is assumed that a considerable number of accesses are made so that information can be obtained. The user who has accessed the shopping mall uses the shopping mall through behavior such as purchase behavior and review transmission (step S11).

  The web server 30 acquires various information transmitted from the user terminal 10. For example, it is assumed that the web server 30 has acquired information related to the evaluation from the user with respect to the shop AAA during the period when the shop AAA operated by the shopkeeper 60 exists. The information related to the evaluation from the user is, for example, information such as a score for the shop AAA transmitted along with the review, a word included in the review, a review score for the product provided in the shop AAA, and a word. is there. Similarly, the web server 30 also acquires information related to the evaluation from the user with respect to the shop BBB operated by the shopkeeper 70.

  In addition, the web server 30 also has attribute information or the like as a business operator of the store exit 60 or the store opener 70 in the shopping mall. For example, the web server 30 uses, as attribute information, the legal characteristics of the store 60 and the store 70 (for example, the types of corporations, limited companies, limited partnerships, individual business owners, foundations, associations, etc.) Information on the types of electricity, communications, manufacturing, food retail, clothing retail, etc.). For example, the web server 30 acquires such attribute information and stores it in the database in association with the identification information of the business operator in the acceptance procedure when the business operator opens a store.

  As described above, using the information acquired by the web server 30, the calculation device 100 calculates the credit rating for the business operator (the store operator 70 in FIG. 1) that opens a store in the shopping mall. In this case, the calculation apparatus 100 acquires information regarding the store exit 60 and the store opener 70 from the web server 30. Specifically, the calculation device 100 acquires information related to the evaluation of the store exit 60 and the store opener 70 by the user using the shopping mall (step S12).

  And the calculation apparatus 100 produces | generates the model which outputs a predetermined index value (score) based on the acquired information (step S13). Specifically, the calculation device 100 is a model generated based on information related to the store opening 60, and when information related to the store opening 70 is input, the correlation between the store closing 60 and the store opening 70 is provided. A model that outputs a score indicating gender is generated. Although details of the generation process will be described later, for example, the calculation device 100 sets the store 60 as correct answer data (teacher data, objective variable), and information on evaluation from the user with respect to the store 60 and attribute information of the store 60 Generate a regression learning model such as Support Vector Machine with features as explanatory variables. And the calculation apparatus 100 outputs the score which shows the correlation with the store closing person 60 which is correct data by inputting the information regarding the shop opening person 70 into the produced | generated model.

  Specifically, the calculation apparatus 100 uses the generated model to calculate a credit rating for the store opening person 70 (step S14). For example, the calculation device 100 assumes that the correlation between the store opening 60 and the store opening 70 is low, that is, the lower the score output from the model is, the lower the possibility that the store opening 70 will close the store is. Calculate the creditworthiness of. On the other hand, since it is assumed that the higher the score output from the model is, the higher the possibility that the store 70 will leave the store, the calculation device 100 calculates the reliability of the store 70 low.

  And the calculation apparatus 100 determines the loan conditions to the store opening person 70 based on the calculated reliability (step S15). Specifically, when the calculated credit rating is higher than a predetermined value, the calculation device 100 expands the credit line or lowers the interest rate compared to the case where the credit rating is lower than the predetermined value. May be determined to be applied. Alternatively, when the calculated credit rating is lower than the predetermined value, the calculation device 100 is applied with a loan condition such as reducing the credit limit or raising the interest rate compared to the case where the credit rating is higher than the predetermined value. It may be determined that.

  As described above, the calculation device 100 according to the embodiment acquires information on the store exit 60 who has lost the right to perform a commercial transaction at a predetermined commercial transaction site on a network such as a shopping mall. And the calculation apparatus 100 calculates the credit rating with respect to the store opening person 70 who opens a store in the shopping mall which is a predetermined | prescribed commercial transaction site based on the information regarding the acquired store closing person 60. FIG.

  In other words, the calculation device 100 according to the embodiment opens a store based on not only financial information of a company that is generally used to measure the creditworthiness of an operator, but also information on the store exit 60 in a commercial transaction site such as a shopping mall. The credit rating of the person 70 is calculated. The store exit 60 in the commercial transaction site has a business form that does not conform to the commercial transaction site, operates unsuitable for the terms of the commercial transaction site, or has an extremely poor evaluation (contents of user reviews) from the user. It is assumed that it is a business operator. The calculation apparatus 100 according to the embodiment treats such information regarding the store 60 as so-called bad data, and refers to the correlation with the bad data, so that the store opener who is currently a store is currently open. A credit rating of 70 is calculated. As a result, the calculation device 100 can calculate a trustworthiness corresponding to the actual situation at the predetermined commercial transaction site for the store opening person 70.

  Moreover, the calculation apparatus 100 can acquire the reaction from the user who uses a shopping mall via a network in real time, and can use it for calculation processing. For this reason, the calculation apparatus 100 can calculate the creditworthiness of the shopkeeper 70 using data with high freshness. In this respect as well, the calculation device 100 can calculate the degree of trust according to the actual situation at a predetermined commercial transaction site.

  Note that the credit level calculated by the calculation device 100 may not necessarily be used for the purpose of financing or investment. For example, the credit level calculated by the calculation device 100 may be used as the credit level as a member store in the shopping mall or as the credit level of the advertiser when the store 70 requests advertisement distribution. Hereinafter, the configuration of the calculation apparatus 100 that performs the above-described processing and the calculation system 1 including the calculation apparatus 100 will be described in detail.

[2. Calculation system configuration)
The configuration of the calculation system 1 including the calculation device 100 according to the embodiment will be described with reference to FIG. FIG. 2 is a diagram illustrating a configuration example of the calculation system 1 according to the embodiment. As illustrated in FIG. 2, the calculation system 1 according to the embodiment includes a user terminal 10, a web server 30, and a calculation device 100. These various apparatuses are communicably connected via a network N (for example, the Internet) in a wired or wireless manner. The calculation system 1 illustrated in FIG. 2 may include a plurality of user terminals 10 and a plurality of web servers 30.

  The user terminal 10 is, for example, an information processing apparatus such as a smartphone, a desktop PC (Personal Computer), a notebook PC, a tablet terminal, a mobile phone, a PDA (Personal Digital Assistant), or a wearable device. It is. The user terminal 10 accesses the web server 30 in accordance with an operation by the user, thereby causing a web page (for example, the web page 61 of the shop AAA shown in FIG. 1 or the web of the shop BBB shown in FIG. 1 to be displayed). Page 71). Then, the user terminal 10 displays the acquired web page on a display device (for example, a liquid crystal display). In the present specification, the user and the user terminal 10 may be identified with each other. For example, “providing information content to the user” may actually mean “providing information content to the user terminal 10 used by the user”.

  The web server 30 is a server device that provides various web pages when accessed from the user terminal 10. In the embodiment, an example has been shown in which the web server 30 provides a website corresponding to a shopping mall and provides a web page of a storekeeper 70 opening a store in the shopping mall. However, the web page provided by the web server 30 is not limited to this, and the web server 30 includes, for example, a news site, a weather forecast site, a shopping site, a finance (stock price) site, a route search site, a map providing site, a travel site, Various web pages related to restaurants introduction sites, web blogs, etc. may be provided.

  Further, the web server 30 may obtain information on each website and information on the network. The information on the network is information transmitted from the user terminal 10 in accordance with a user operation when using services provided from various websites, for example. For example, the information on the network includes information related to a search query in a search site, information related to purchase behavior in a shopping site and reviews from a user, information related to a message in an SNS (Social Networking Service) site posted by the user, and the like.

  The calculation device 100 acquires information about the store exiter 60 who has lost the right to conduct a commercial transaction at a predetermined commercial transaction site, and calculates a credit rating for the store opener 70 opening a store at the predetermined commercial transaction site based on the acquired information Device. The calculation device 100 may also serve as the configuration as the web server 30 described above. That is, the calculation apparatus 100 itself may provide a shopping mall, and the calculation apparatus 100 itself may acquire information from the user terminal 10 that has accessed the shopping mall.

  In addition, the calculation device 100 is an information processing terminal installed in a financial institution that actually provides financing to a business operator, or information installed in a financial management company that conducts research, acquisition, management, and provision of financial information related to the business operator. You may mutually communicate with a processing terminal etc. For example, the calculation apparatus 100 may acquire information regarding determination of loan conditions from an information processing terminal installed in a financial institution or the like. Further, when the financial information of the store 60 and the store 70 is published, the calculation device 100 may appropriately acquire financial information from an information processing terminal installed in a financial management company or the like.

[3. Configuration of calculation device]
Next, the configuration of the calculation apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 3 is a diagram illustrating a configuration example of the calculation apparatus 100 according to the embodiment. As illustrated in FIG. 3, the calculation device 100 includes a communication unit 110, a storage unit 120, and a control unit 130. The calculation device 100 includes an input unit (for example, a keyboard and a mouse) that receives various operations from an administrator who uses the calculation device 100, and a display unit (for example, a liquid crystal display) that displays various types of information. You may have.

(About the communication unit 110)
The communication unit 110 is realized by, for example, a NIC (Network Interface Card). The communication unit 110 is connected to a communication network by wire or wireless, and transmits and receives information to and from the user terminal 10 and the like via the communication network.

(About the 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. The storage unit 120 according to the embodiment includes a store opening information storage unit 121, a feature information storage unit 122, a model storage unit 127, and a loan condition storage unit 128. Hereinafter, each storage unit will be described in order.

(Regarding store information storage unit 121)
Store opening information storage part 121 memorizes information about a business opening a store on a predetermined commercial transaction site. For example, the information stored in the store opening information storage unit 121 is acquired from the web server 30 via the communication network. Here, FIG. 4 shows an example of the store opening information storage unit 121 according to the embodiment. As shown in FIG. 4, the store opening information storage unit 121 stores “store opening”, “store account”, “store opening date”, “store opening date”, “reason for closing”, “identity information”, “credit”. It has items such as. The item of “feature information” includes small items such as “business information”, “store information”, “product information”, “word information”, and “financial information”.

  “Store opener” indicates the name of a business operator that has opened a store in the past or at the present time. The “store account” indicates an account issued from the web server 30 side when a business operator opens a store on a commercial transaction site. In addition, as shown in FIG. 4, the business operator can have a plurality of store accounts. For example, the business operator can use the store account properly according to the category of the product to be handled and the business form. In addition, when the web server 30 disposes of a store forcibly with respect to one store account among store accounts owned by the business, the web server 30 also terminates other store accounts owned by the same business (right to open a store). May be performed).

  “Store opening date” indicates the date when the store opened on the commerce site. “Closed date” indicates the date when the store is closed from the commerce site. For a business operator (or store account) that has not left the store at this time, the item of the store closing date is blank ("-"). “Reason for leaving” indicates the reason for leaving the store. The reason for leaving the store is used, for example, to distinguish correct data in model generation processing described later. For example, “forced exit” is an account forcibly exited from the web server 30 side, and therefore, it is assumed that there is a high possibility that the data is unsatisfactory. On the other hand, “Closed” is a voluntary closing by a business operator, so it is assumed that there is a low possibility that the data is unsatisfactory.

  “Feature information” indicates information that can be a feature in model generation processing (modeling) by the calculation apparatus 100. “Business information” indicates, for example, attribute information of a business operator. “Store information” indicates, for example, a rating from a user to a store operated by a business operator. “Product information” indicates, for example, a score from a user of a product handled in a store operated by a business operator. “Word information” indicates information about a word included in a review posted from the user to the store. “Financial information” indicates financial information of a business operator that is publicized or surveyed by the above-mentioned financial management company or the like.

  In the example shown in FIG. 4, feature information such as “business information” is indicated by a concept such as “aaa”. It is assumed that information such as the weight value of each attribute information calculated in the process is stored. Details of each feature information will be described with reference to FIGS.

  “Credit level” indicates the credit level of the business. The credit rating is stored in the store information storage unit 121 after the credit level is calculated by the calculation unit 134 described later. As shown in FIG. 4, for example, for a store exit (store account) that has already left the store, the item of credit rating is blank.

  In other words, FIG. 4 shows that “store branch 60” has store accounts A01 and A02 as an example of data stored in the store opening information storage unit 121. In addition, the store corresponding to the store account A01 has a store opening date of “July 1, 2014”, a store closing date of “March 1, 2016”, and the reason for leaving the store is “Forced closing”. is there. The feature information of the store account A01 includes business information “aaa”, store information “bbb”, product information “ccc”, word information “ddd”, and financial information “ eee ". Further, since the store account A01 is closed, the credit rating is not calculated. On the other hand, FIG. 4 shows that the credit rating of the store account A11 is calculated as “75”.

(About the feature information storage unit 122)
The feature information storage unit 122 stores information about features used in the model generation process. Specifically, the feature information storage unit 122 stores information on evaluations from users with respect to store exits and store openings on a commercial transaction site, attribute information of store exits and store openings, and the like. As shown in FIG. 3, the feature information storage unit 122 includes data tables such as a business information table 123, a store information table 124, a product information table 125, and a word information table 126. Hereinafter, each data table will be described in order.

(About business information table 123)
The business information table 123 stores attribute information and the like as business operators of store exits and store openings. Here, FIG. 5 shows an example of the business information table 123 according to the embodiment. As shown in FIG. 5, the business information table 123 includes items such as “store branch”, “store account”, “corporate personality”, “business type”, and “continuation years”.

  “Store opening” and “store account” correspond to the same items stored in the store opening information storage unit 121. “Corporate personality” indicates the legal personality of the store operator. “Industry” indicates the type of store. The business type may correspond to, for example, a product category handled in the store. In addition, one type of industry is not set in one store, but a plurality of types of industry may be set. “Continuation years” indicates the number of years that the store has continued.

  That is, in FIG. 5, as an example of data stored in the business information table 123, the legality of the store 60 is “individual business owner”, the business type of the store account A01 is “fashion”, and the store account A02 The business type is “food”, the store account A01 has a duration of “1 year and 8 months”, and the store account A02 has a duration of “0 years and 7 months”.

(About the store information table 124)
The store information table 124 stores information related to a store (store) operated by the store operator. Here, FIG. 6 shows an example of the store information table 124 according to the embodiment. As shown in FIG. 6, the store information table 124 includes “store branch”, “store account”, “number of sales”, “sales amount”, “PV (1 day)”, “number of reviews”, and “store score”. It has items such as.

  “Store opening” and “store account” correspond to the same items stored in the store opening information storage unit 121. “Number of sales” indicates the number of sales of products in the store. “Sales amount” indicates the sales amount of the product in the store. In the example shown in FIG. 6, the information stored in “number of sales” and “sales amount” is indicated by concepts such as “B01” and “C01”. The item “Sales Amount” includes various information related to the number of sales and the amount of sales, such as the number of sales of specific products, the value of sales, the ratio of sales to the number of products handled, and the period of counting the number of items and the amount of money. Shall be stored.

  “PV (1 day)” indicates the number of accesses (page view) per day when the user accesses the shop page of the store corresponding to the store account in the commercial transaction site. For example, in the PV item, the number of PVs aggregated every day may be stored as needed, or the number of PVs per day when the total number of accesses in a predetermined period is averaged may be stored.

  “Number of reviews” indicates the number of reviews submitted by the user to the store. In the example shown in FIG. 6, the average number of reviews per day in the store is stored as the number of reviews. “Store score” indicates an index value for evaluating a store in a review posted by a user. For example, the store score is indicated by a numerical value in five stages from 1 to 5, where 5 is the highest and 1 is the lowest. In the example shown in FIG. 6, an example is shown in which the average score of the reviews in the review posted by the user before the store leaves is stored as the store score. Note that the number of reviews and the store score may be stored as an average of the numbers and scores calculated in a predetermined period (for example, one week), the cumulative number up to the present time may be stored, or the cumulative number up to the current time may be stored. Information aggregated in various manners may be stored, such as the number divided by the number of years in the store.

  That is, in FIG. 6, as an example of data stored in the store information table 124, the number of sales of the store account A01 that the store 60 has is “B01”, the sales amount is “C01”, and the PV of the store (1 day) indicates “350” times, the average number of reviews per day is “20”, and the store score is “3.4”.

  Although illustration in FIG. 6 is omitted, the store information table 124 may store various information related to the store in addition to the information described above. For example, in the store information table 124, the number of users who continue to use the store (for example, users who perform an action such as accessing again within one week) and the user unit price of the user (for example, 1 in a predetermined period). Information such as purchase amount per user) may be stored.

(About the product information table 125)
The merchandise information table 125 stores information related to merchandise provided by the shopkeeper in the store. Here, FIG. 7 shows an example of the product information table 125 according to the embodiment. As illustrated in FIG. 7, the product information table 125 includes items such as “store branch”, “store account”, “product ID”, “product score”, and “total number of reviews”.

  “Store opening” and “store account” correspond to the same items stored in the store opening information storage unit 121. “Product ID” indicates identification information for identifying a product handled in the store. “Product score” indicates a score in a review posted by a user for a product. For example, like the store score shown in FIG. 6, the product score is indicated by a numerical value of 5 levels from 1 to 5, where 5 is the highest evaluation and 1 is the lowest evaluation. In the example shown in FIG. 7, an example is shown in which the average score of the reviews in the review posted by the user until the handling of the product in the store is completed is stored as the product score. “Total number of reviews” indicates the total number of reviews submitted for the product.

  That is, in FIG. 7, as an example of data stored in the product information table 125, the product score of the product P01 identified by the product ID “P01” handled by the store account A01 possessed by the store 60 is “3.2. ”Indicating that the total number of reviews is“ 125 ”.

  Although illustration in FIG. 7 is omitted, the product information table 125 includes information on the period in which reviews are tabulated, the ratio of the number of reviews posted to the PV, the number of sales and the sales amount for each product. Such information may be stored.

(Regarding the word information table 126)
The word information table 126 stores information about words included in the text data posted from the user as a review to the store owner (store). Here, FIG. 8 shows an example of the word information table 126 according to the embodiment. As shown in FIG. 8, the word information table 126 has items such as “store branch”, “store account”, “appearance word”, and “word appearance count”.

  “Store opening” and “store account” correspond to the same items stored in the store opening information storage unit 121. “Appearance word” indicates a word that appears in a review posted by a user for a store or a product. The “word appearance count” indicates the number of times each word appears in the aggregated text data.

  That is, in FIG. 8, as an example of data stored in the word information table 126, in the review posted to the store corresponding to the store account A01 that the store exit 60 has, “product” is “80” times, “Fake” appears “51” times, “Overseas” appears “26” times, “Take” appears “16” times, “Closed” appears “12” times, and “Worst” appears “9” times ing. Further, in the review posted to the store corresponding to the store account A11 that the store 70 has, “packing” is “57” times, “delivery” is “50” times, “quick” is “46” times, This shows that “shipping” appears “31” times, “next day” appears “25” times, and “help” appears “14” times.

  In FIG. 8, the number of word appearances is shown as the number of times, but the information corresponding to the number of word appearances may be information such as a ratio of the total number of words included in the text data, for example. Moreover, the information memorize | stored in the word information table 126 may memorize | store the information totaled for every predetermined period. For example, it is predicted that the reviews from the users will differ in the content posted every season. For this reason, the word information table 126 may store, for example, information comparing words appearing in the text data aggregated at the same time, the number of appearances, and the like.

  Note that the calculation apparatus 100 may perform morphological analysis on text data posted by the user, for example, when tabulating information about words. Furthermore, the calculation apparatus 100 may select in advance the words to be counted based on the word weights in one review. For example, the calculation apparatus 100 may store only words having high weights using an algorithm that scores each word appearing in the review, such as tf-idf (Term Frequency-Inverse Document Frequency). Further, the calculation device 100 may store all words that appear in the review regardless of the number of appearances, weights, and the like.

(About the model storage unit 127)
The model storage unit 127 stores information related to the model generated by the calculation device 100. Here, FIG. 9 shows an example of the model storage unit 127 according to the embodiment. As illustrated in FIG. 9, the model storage unit 127 includes items such as “model ID”, “information update date”, and “correct answer data”.

  “Model ID” indicates identification information for identifying a model. “Information update date” indicates the date on which the model was updated. “Correct answer data” indicates a case where the generated model is correct data. For example, a model in which the item of correct answer data is “forced store clerk” indicates that the data corresponding to the store account whose reason for leaving the store is “forced store close” is learned as correct data.

  That is, in FIG. 9, as an example of data stored in the model storage unit 127, the model M01 identified by the model ID “M01” is updated to “November 1, 2016”, and the model M01 The case where is the correct answer data indicates that the store account corresponds to the “forced store quit”.

(About loan condition storage unit 128)
The loan condition storage unit 128 stores information regarding loan conditions determined based on the calculated creditworthiness. Here, FIG. 10 illustrates an example of the loan condition storage unit 128 according to the embodiment. As illustrated in FIG. 10, the loan condition storage unit 128 includes items such as “condition ID”, “credit rating”, and “loan condition”.

  “Condition ID” indicates identification information for identifying a condition. “Trustworthiness” indicates the trustworthiness with respect to the store clerk calculated by the calculation device 100. The loan condition indicates the content of the loan condition determined by the calculation device 100. In the example shown in FIG. 10, the loan condition is indicated by a concept such as “H01”. Actually, the items of the loan condition include a credit line (amount available for loan), an interest rate, and repayment. It is assumed that specific conditions such as a period are stored.

  That is, in FIG. 10, as an example of data stored in the loan condition storage unit 128, the loan condition G01 identified by the condition ID “G01” is for the shopkeepers whose credit ratings are “100” to “81”. The loan condition is “H01”.

(About the control unit 130)
The control unit 130 is a controller, for example, various programs (an example of a calculation program) stored in a storage device inside the calculation device 100 by a CPU (Central Processing Unit), an MPU (Micro Processing Unit), or the like. This is realized by executing the RAM as a work area. The control unit 130 is a controller, and is realized by an integrated circuit such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  As shown in FIG. 3, the control unit 130 according to the embodiment includes an acquisition unit 131, a generation unit 132, a reception unit 133, a calculation unit 134, and a determination unit 135, and information processing described below Realize or execute the functions and operations of Note that the internal configuration of the control unit 130 is not limited to the configuration illustrated in FIG. 3, and may be another configuration as long as the information processing described below is performed. Further, the connection relationship between the processing units included in the control unit 130 is not limited to the connection relationship illustrated in FIG. 3, and may be another connection relationship.

(About the acquisition unit 131)
The acquisition unit 131 acquires information regarding the store exiter 60 who has lost the right to perform a commercial transaction at a predetermined commercial transaction site on the network. For example, the acquisition unit 131 acquires information related to an evaluation from the user with respect to the store exit 60 on a predetermined commercial transaction site as information about the store exit 60.

  The acquisition unit 131 uses the store (store) where the store 60 has opened or the number of reactions from the user with respect to the product, or the store where the store 60 has opened as information on the evaluation of the store 60 by the user. Information on evaluation values scored by the user for the product is acquired. Specifically, the acquisition unit 131 acquires information related to reviews from the user stored in the feature information storage unit 122.

  Further, the acquisition unit 131 changes the number of reactions from the user to the store or product opened by the store 60, or the evaluation value evaluated by the user for the store or product opened by the store 60 You may acquire the information about. That is, the acquisition unit 131 acquires an increase / decrease in the number of reviews posted in a predetermined period and a transition in an increase / decrease in scores. Thereby, the acquisition part 131 can acquire the information regarding the tendency that the store which received high evaluation from the user has received low evaluation in recent years. In addition, among businesses that open a store on a commercial transaction site, for example, there may be a business that intends to bankrupt a store systematically based on a financial strategy. For example, for such a business operator, a phenomenon may be observed in which the evaluation from the user is suddenly reduced before the sales of the product are reduced. The acquisition unit 131 can acquire information for determining such an operator by acquiring a transition of information regarding evaluation.

  Moreover, the acquisition part 131 acquires the word contained in the review from the user with respect to the store or goods which the store opening person 60 opened as information regarding the evaluation from the user with respect to the store closing person 60. FIG. For example, the acquisition unit 131 performs morphological analysis on reviews from the user, extracts words included in the reviews, and acquires information about the extracted words. Further, the acquisition unit 131 may acquire only words having high weights using an algorithm such as tf-idf that scores each word appearing in the review.

  Moreover, the acquisition part 131 acquires the attribute information as a company of the storekeeper 60 in a predetermined | prescribed commercial transaction site as information regarding the storekeeper 60. FIG. Specifically, the acquisition unit 131 acquires, as attribute information, a legal personality as a business operator of the store exit 60 or a business type as a business operator of the store exit 60. The industry in this case is not an industry in a narrow sense such as manufacturing industry or communication industry. For example, it is an industry in a broad sense that shows the characteristics of the store as indicated by the category of products handled in the store where the store opened. Meaning may be included.

  In addition, the acquisition unit 131 acquires the reason for leaving the store in each store account of the store 60. For example, a reason for a store clerk to leave a store on a commercial transaction site is a forced store exit from the site side, a store exit that becomes a companion to a store account that has been forcibly closed, or a request from the store owner side It may be closed or closed. The acquisition unit 131 acquires these reasons as information related to the store closing person 60 and stores them in the store opening information storage unit 121 in association with the store account of the store closing person 60.

  Moreover, the acquisition part 131 memorize | stores the acquired information in each memory | storage part suitably. For example, the acquisition unit 131 stores the acquired business operator attribute information and the like in the business information table 123. Further, the acquisition unit 131 stores the acquired information related to the evaluation from the user in the store information table 124, the product information table 125, and the word information table 126. In addition, when the acquisition unit 131 acquires a loan condition set by, for example, an administrator of the calculation apparatus 100 or a financial institution, the acquisition unit 131 stores information on the loan condition in the loan condition storage unit 128. In addition, the acquisition unit 131 acquires not only information related to the store opening 60 but also information related to the store opening 70 that is a target of credit calculation.

  Note that the acquisition unit 131 may randomly acquire information on various store exits 60 and store openings 70 without specifying the store exits 60 and store openings 70 from which information is acquired. For example, the acquisition unit 131 acquires information about the store opening 60 and the store opening 70 at any time by using a program such as a search robot used for a search engine or the like to crawl on the Internet, or the acquired information Or may be updated.

  Further, the acquisition unit 131 may acquire financial information when the financial information of the store exit 60 and the store opener 70 can be acquired from, for example, a financial institution or a financial management company. When the acquisition unit 131 acquires financial information, the calculation unit 134 to be described later may calculate the credit rating for the storekeeper 70 in consideration of the financial information.

(About the generator 132)
The generation unit 132 is a model that is generated based on information about the store opening 60, and when information about the store opening 70 is input, a score indicating a correlation between the store closing 60 and the store opening 70 is obtained. Generate output model.

  The production | generation part 132 produces | generates a model based on the information regarding the evaluation from the user with respect to the shopkeeper 60 in a predetermined | prescribed commercial transaction site. For example, the generation unit 132 correlates a word included in a review from a user for a store or a product opened by the store 60 and a word included in a review from the user for a store or a product opened by the store 70. A model that outputs a score based on the above may be generated.

  Moreover, the production | generation part 132 may produce | generate a model based on the attribute information as a company of the storekeeper 60 in a predetermined | prescribed commercial transaction site. For example, the generation unit 132 may have a legal personality as the business operator of the store opening 60 and a corporate personality as the business operator of the store opening 70, or a business type as the business operator of the closing store 60 and a business type as the business operator of the store opening 70. A model that outputs a score may be generated based on the correlation with the.

  As described above, the generation unit 132 uses each piece of information regarding the store exit 60 acquired by the acquisition unit 131 as a feature, and generates a model for guessing the store exit 60 that is correct data. Specifically, the generation unit 132 generates a model that recursively learns what kind of information is given what weight among each piece of information regarding the store 60 that becomes the feature. In other words, the generation unit 132 indicates which of the pieces of information related to the store exit 60 acquired as correct answer data has an influence on an unfavorable result of “store exit”. Generate a model.

  Below, an example of the model which the production | generation part 132 produces | generates is shown. Note that the model generated by the generation unit 132 is not limited to the following model, and is a model generated based on information about the store opening 60, and when the information about the store opening 70 is input, the store closing 60 Any model may be used as long as it is a model that outputs a score indicating the correlation between the shop and the store 70.

  For example, the generation unit 132 sets the result information indicating that the store exit 60 has ended at the present time as an objective variable in machine learning. And the production | generation part 132 makes each information regarding the store exit 60 acquired while the store exit 60 opened the store an explanatory variable (feature). And the production | generation part 132 produces | generates the model regarding the shopkeeper 60 using an objective variable and an explanatory variable. The generation unit 132 generates a model using various explanatory variables. For example, the generation unit 132 can use any information stored in the feature information storage unit 122 as an explanatory variable.

  For example, the generating unit 132 generates a result information indicating whether or not the store 60 has left the store and an expression indicating information about the store 60. Furthermore, the generation unit 132 learns and calculates what weight each piece of information about the store closing person 60 has for the event that the store closing person 60 has left the store. Thereby, the production | generation part 132 can obtain the information of how much each information contributes with respect to the phenomenon of leaving a store. For example, the generation unit 132 creates the following formula (1).

y = ω ₁ · x ₁ + ω ₂ · x ₂ + ω ₃ · x ₃ ... + ω _N · x _N (1) (N is an arbitrary number)

  The above formula (1) is created for each individual store exit 60, for example. In the above formula (1), “y” indicates an event “whether or not the store 60 has lost the right to open a store on the commerce site (whether or not it has left)”. In learning, if the store exit 60 is to be correct data, “y” takes a value of “1”. In addition, in learning, if the store 70 who has not left the store at the present time is used as correct answer data, “y” takes a value of “0”. Note that, as in the embodiment, if the determination of the store exit is performed for each store account, the value substituted for “y” corresponds to the store exit result for each store account.

In the above formula (1), “x” corresponds to each explanatory variable that is information (features) about the shopkeeper 60. In the above formula (1), “ω” is a coefficient of “x” and indicates a predetermined weight value. Specifically, “ω ₁ ” is a weight value of “x ₁ ”, “ω ₂ ” is a weight value of “x ₂ ”, and “ω ₃ ” is a weight value of “x ₃ ”. It is. Thus, the above equation (1) is a variable (for example, “ω ₁ ” including the explanatory variable “x” corresponding to the feature information extracted from the information related to the shopkeeper 60 and the predetermined weight value “ω”. -X1 _" ) is created by combining.

For example, in the above formula (1), suppose that “x ₁ ” is “a store score in the store account A01 (which the store exit 60 has)”. Further, “x ₂ ” is assumed to be “corporate personality of the store exit 60”. Further, “x ₃ ” is assumed to be “the number of reviews to the store in the store account A01 (which the store 60 has)”. In this case, the above formula (1) corresponding to the store exit 60 can be expressed as the following formula (2).

y _{(store 60)} (= 1) = ω ₁ · (store score in store account A01 (= 3.4)) + ω ₂ · (corporate personality (= individual business owner)) + ω ₃ · (store account A01 Number of reviews for the store (= 20))) + ω _N · x _N (2)

The generating unit 132 generates an expression for each store account 60, more specifically, for each store account of the store exit 60, as in the above formula (2). Then, the generation unit 132 sets the generated expression as a machine learning sample. Then, the generation unit 132 derives a value corresponding to the predetermined weight value “ω” by performing calculation processing (regression learning) of a sample expression. That is, the generation unit 132 determines a predetermined weight value “ω” that satisfies the above equation (2). In other words, the generation unit 132 can determine the weight value “ω” indicating the influence of the predetermined explanatory variable on the objective variable “y”. For example, if “corporate personality” contributes significantly to the event “exit” compared to other variables, the value of the weight value “ω ₂ ” corresponding to “corporate personality” A large value is calculated in comparison with the variable.

  In the above example, the store score and the number of reviews are shown as numerical values. However, these numerical values may be converted to be used as explanatory variables by a predetermined setting. For example, the generation unit 132 may normalize and use these numerical values by a known method. Alternatively, for example, the generation unit 132 may divide the store score into three stages of “high”, “medium”, and “low”, and use each as an explanatory variable. In this case, for example, the generation unit 132 can calculate the weight value given to the objective variable (= store opening) by the explanatory variable that the store score is “low”. Specifically, as a result of learning, the generation unit 132 can learn that a store opening person with a lower store score has a higher possibility of closing a store.

  As described above, the generation unit 132 generates a model for determining the relevance between the tendency that a business that has opened a store on a commercial transaction site reaches “store exit” and the information (feature) about the business. To do. That is, when the information of the store opening 70 to be processed is input to the generated model, the generating unit 132 shows a tendency that the information is “exit”, in other words, how similar to the store closing 60. It is possible to output a score indicating the relevance such as having. In the calculation process using the above equation (2), the left side is not set to “1” or “0” itself, but a predetermined error is assumed, and a value obtained by squaring the difference from the error is a minimum value. The optimal solution of “ω” may be calculated using a method such as a least square method that approximates to

  Although the notation is omitted in the above example, the generation unit 132 may generate a model using all the features related to the store exit 60. For example, the generation unit 132 may generate a model using any word posted from the user to the store clerk 60 as an explanatory variable. Thereby, for example, the generation unit 132 has a high tendency to leave a store where “fake”, “overseas”, and “taken” are included in the review, and “packing”, “delivery”, and “rapid” are included in the review. There may be a case where a model that finds a tendency that a store that contains “has a low tendency to leave” may be generated. This means that, when the word “fake” is included in the evaluation from the user for the store 70 to be processed, the store 70 is likely to leave the store in the future. In other words, it can be said that the word “fake” is an element having a heavy weight (as a feature for obtaining the correlation with the store exit 60) for the result of “exit”.

  In addition, the generation part 132 may consider the appearance number of a word, when using a word as a feature. For example, the generation unit 132 may perform an adjustment such that the weight calculated for a predetermined word is further multiplied by a predetermined ratio according to the number of times the word appears. As a result, the generation unit 132 outputs a score indicating a stronger correlation with the store closing person 60 for the store opening person 70 in which a word indicating a strong correlation with the store closing person 60 frequently appears. Can be generated. Further, the generation unit 132 uses a weight such as tf-idf in a predetermined document (one review submitted by a user, one review for a product, etc.) for a weight calculated for a predetermined word. Further, an adjustment may be made such that a predetermined ratio is multiplied.

  In addition, after such learning, the generation unit 132 tends to indicate that, for example, “individual business owners” among corporations have a high possibility of leaving, and “corporations” and “foundations” have a low possibility of leaving. May be derived. Also in this case, it can be said that the attribute “individual business owner” has a higher weight than the attribute “corporation” or “foundation” with respect to the result of “leaving the store”.

  Note that the generation unit 132 may generate a model in consideration of various information. For example, the generation unit 132 may include a credit calculated from a known model when generating the model of the store exit 60 as described above.

  As an example, the generation unit 132 may use any known company evaluation model as the existing model. For example, the Alt value (Edward Altman) Z value and the like are widely known as a company evaluation model for obtaining a score indicating the probability of bankruptcy of an operator based on financial information. Using such an existing model, the generation unit 132 uses a score indicating the bankruptcy probability of the business operator (in the embodiment, it may be read as “probability that the store opens from the commercial transaction site”) as an explanatory variable 1 Alternatively, it may be incorporated into the above model.

  The model illustrated is an example, and the model generated by the generation unit 132 is not limited to the above example. That is, the generation unit 132 is a model that is generated based on information related to the store opening 60, and shows the correlation between the store closing 60 and the store opening 70 when information related to the store opening 70 is input. If it is a model that outputs a score, a model different from the above example may be appropriately generated by combining known methods. For example, in the above example, an example in which a model is generated by regression analysis has been shown. However, the generation unit 132 may generate a model by other statistical processing, or a model using a neural network or the like. It may be generated.

  Moreover, the production | generation part 132 may distinguish correct data for every classification of reason for leaving a store. Generally, compared to the store exit 60 that has exited due to voluntary circumstances, the store exit 60 that is forcibly exited from the web server 30 side is assumed to be an operator with a lower credit rating. For this reason, the generation unit 132 may generate a model that narrows down only the store account that has been “forced to leave”, for example, the store 60 to be correct data. Alternatively, the generation unit 132 may generate a model that narrows down the store 60 to be correct data only to store accounts of “(voluntary)” or “closed”. Thereby, the generation unit 132 has a score indicating whether the store 70 is a business that is predicted to be simply “leave a store” or a business that is highly likely to be “forced to leave”. Can be generated separately. In other words, the generation unit 132 improves the accuracy of outputting a score indicating a certain event (whether the store is simply closed or forced to close) by generating different models by properly using the correct answer data and features to be generated. A model can be generated.

(About the reception unit 133)
The accepting unit 133 accepts a request related to the creditworthiness of the shopkeeper 70. Specifically, the accepting unit 133 according to the embodiment accepts a request for calculating the creditworthiness of the shopkeeper 70 from an information processing terminal such as a financial institution that intends to provide a loan. The accepting unit 133 may cause the obtaining unit 131 to obtain information on the store seller 70 to be processed, triggered by accepting the request. For example, the reception unit 133 causes the user's evaluation of the store 70 to store, the attribute information of the store 70, and the like. Then, the reception unit 133 sends the received information to the generation unit 132 and the calculation unit 134.

(About the calculation unit 134)
The calculation unit 134 calculates the credit rating of the store opening person 70 who opens a store on a predetermined commercial transaction site based on the information regarding the store closing person 60 acquired by the acquiring unit 131. For example, the calculation unit 134 calculates the creditworthiness of the store seller 70 based on the score output from the model generated by the generation unit 132.

  Specifically, the calculation unit 134 acquires the score of the store opening 70 that is the processing target to be output by inputting information on the store opening 70 to be processed into the model generated by the generation unit 132. The output score indicates the correlation between the store opening person 70 and the shop closing person 60 to be processed. For example, an event that the correlation between the store opening person 70 and the store opening person 60 to be processed is high indicates that the shop opening person 70 is highly likely to leave the store. The calculation unit 134 calculates the creditworthiness of the shopkeeper 70 based on the score. The calculating unit 134 may use the output score itself as the creditworthiness of the storekeeper 70, or may perform processing such as normalizing the output score, and may use the value calculated after the processing as the creditworthiness of the storekeeper 70. . For example, the calculation unit 134 may perform a process of normalizing the score output from the model to a numerical value of 100 levels, and the value from 1 to 100 calculated after the process may be used as the creditworthiness of the store seller 70.

  Note that the calculation unit 134 does not necessarily have to calculate the creditworthiness of the shopkeeper 70 using the model generated by the generation unit 132. For example, the calculation unit 134 statistically compares the information about the evaluation of the store opening 70 from the user with the information about the evaluation of the closing store 60 from the user, and determines the creditworthiness of the store opening 70 based on the compared information. It may be calculated.

(About determination unit 135)
The determination unit 135 determines the loan conditions for the store opening person 70 based on the trustworthiness calculated by the calculation unit 134. For example, the determination unit 135 determines which of the conditions held in the loan condition storage unit 128 the credit level calculated by the calculation unit 134 corresponds to determine the loan condition for the store opening person 70.

  The determination unit 135 may notify the storekeeper 70 of the determined loan conditions. Further, the determination unit 135 may notify the financial institution or the like of information that associates the store opening 70 with the loan conditions determined for the store opening 70.

[4. Processing procedure)
Next, a generation process procedure performed by the calculation apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 11 is a flowchart illustrating a generation processing procedure performed by the calculation apparatus 100 according to the embodiment.

  As illustrated in FIG. 11, the acquisition unit 131 acquires information regarding the store exit 60 from the web server 30 (step S101). And the production | generation part 132 is based on the information acquired by the acquisition part 131, and the model which outputs the score which shows the correlation with the store opening person 60 and the shop opening person 70, when the information regarding the shop opening person 70 is input. Is generated (step S102). The generation unit 132 stores the generated model in the model storage unit 127 (step S103).

  Next, the procedure of the calculation process performed by the calculation apparatus 100 according to the embodiment will be described with reference to FIG. FIG. 12 is a flowchart illustrating a calculation processing procedure performed by the calculation apparatus 100 according to the embodiment.

  As illustrated in FIG. 12, when the request for the credit calculation is received by the reception unit 133, the acquisition unit 131 determines whether or not the information regarding the shopkeeper 70 to be processed has been acquired (Step S <b> 201). . If the acquisition unit 131 has not acquired information, the acquisition unit 131 waits until acquisition (step S201; No).

  On the other hand, when the information regarding the store opening 70 is acquired (step S201; Yes), the calculation unit 134 inputs the acquired information regarding the store opening 70 into the model generated by the generation unit 132 (step S202).

  And the calculation part 134 outputs the score which shows the correlation with the store opening person 70 made into a process target, and the store closing person 60 (step S203). Furthermore, the calculation unit 134 calculates the creditworthiness of the store seller 70 based on the output score (step S204). The determination unit 135 determines the loan conditions for the shopkeeper 70 based on the calculated creditworthiness (step S205).

[5. (Modification)
The calculation device 100 according to the above-described embodiment may be implemented in various different forms other than the above-described embodiment. Therefore, in the following, another embodiment of the calculation device 100 will be described.

[5-1. (Co-occurring words)
In the above embodiment, the calculation apparatus 100 has shown an example in which a word is used as a feature. In such an example, the calculation apparatus 100 may use a pair of a certain word and a word that co-occurs as the feature instead of using one word as a feature.

  For example, when the word “packing” appears in a review from the user, the words “packing” and “poor” are used together, and the words “packing” and “miscellaneous” are used together. It is assumed that there is a gap in the user's evaluation from the case where it is used. For this reason, the calculation apparatus 100 may set a combination that co-occurs for a predetermined word in advance, and may use the set combination of words as a feature.

  Note that the calculation device 100 may store preset data for co-occurring word combinations, or may extract word pairs to be combined by learning processing. Thereby, the calculation apparatus 100 can perform the calculation process of the reliability which reflected the evaluation from the user with higher accuracy.

[5-2. Data to be input to model)
In embodiment mentioned above, the calculation apparatus 100 showed the example which calculates the score which inputs the information regarding the store opening person 70, and shows the correlation with the store closing person 60. FIG. Here, the calculation device 100 may associate the acquisition time of the information on the store clerk 60 used to generate the model with the information on the store clerk 70 input to the model.

  Since the information acquired by the calculation device 100 is information on a commercial transaction site such as a shopping mall, there is a possibility that the information is affected by time. For example, there may be a difference in the number of sales and the sales amount for each store between the time when transactions are actively performed, such as the end of the year, and the other time. For this reason, the calculation apparatus 100 stores the time when information is acquired in order to generate a model. Then, when inputting information on the storekeeper 70 to be processed, the calculation device 100 inputs information acquired at a time similar to the time when the information used for generating the model is acquired. Thereby, the calculation apparatus 100 can perform the calculation process which considered the influence of the time of the output score.

[5-3. Amount of information)
In the above embodiment, the calculation apparatus 100 has shown an example in which a model is generated based on various information that can be acquired on the network. Here, the calculation apparatus 100 may perform processing using only information acquired from general users who use various websites that exceeds a certain threshold.

  For example, reviews regarding products in a certain store, user evaluations, and the like may show a biased tendency if they are not based on a certain number or more of data. In this case, since the influence of the biased data tends to be applied to the regression equation, the calculation device 100 may generate a model that cannot accurately calculate the creditworthiness of the store seller 70. For this reason, the calculation device 100 can use only data for which the total number of scores and reviews transmitted from the user exceeds a certain number for a predetermined store account as data handled in the model calculation process. Thereby, the calculation apparatus 100 can generate a model that outputs a highly reliable score.

[5-4. correction〕
The calculation apparatus 100 may generate a model that can correct the output score according to the actual economic state. For example, the calculation device 100 classifies the store opening 70 whose management state is improved when the yen is strong, the store opening 70 having no influence, and the store opening 70 whose management state deteriorates. Then, when generating the model, the calculation device 100 generates a model that outputs a corrected score in consideration of the trend of the value of the circle by using the trend of the value of the circle in a predetermined period as a feature. To do. For example, the calculation device 100 acquires information on the store 60 and the store 70 in the long term and accumulates information on the linkage with the trend of the value of the yen, thereby using such correction information as a model. To reflect.

[5-5. Information variations)
In the above-described embodiment, the calculation apparatus 100 illustrates an example in which processing is performed using various types of information stored in the feature information storage unit 122. Here, the variation of information handled by the calculation apparatus 100 will be described in more detail.

  For example, the calculation device 100 may acquire more detailed information regarding the words stored in the word information table 126. For example, the calculation device 100 may perform morphological analysis on the acquired word and acquire information aggregated for each part of speech. Specifically, the calculation apparatus 100 acquires a totaling result obtained by classifying noun categories into people, things, places, and the like. In addition, the calculation device 100 can acquire a classification result such as whether each word has a positive attribute or a negative attribute, and a determination result such as an intention determination of the dialog when the message is in an interactive format.

  Moreover, the calculation apparatus 100 may acquire voice information transmitted from the user as a message. That is, the calculation apparatus 100 may acquire the text information including the word by recognizing the voice transmitted from the user who uses voice search or voice input.

  Moreover, the calculation apparatus 100 may extract only a specific part of speech regarding the acquired word. For example, the calculation device 100 may use a part of speech that can represent the situation of the store 60 and the store 70, such as a noun and an adjective, in the process. Thereby, the calculation apparatus 100 can reduce the processing load without reducing the accuracy of score calculation.

  Moreover, the calculation apparatus 100 may use only words registered in advance for the score calculation process. For example, the calculation apparatus 100 may use a word that can express the evaluation of the store clerk 60 and the store clerk 70 such as “good” or “bad” in the process. Such registration is performed by an administrator of the calculation apparatus 100, for example.

  In addition to classifying words as parts of speech, the calculation device 100 may classify the words for each predetermined category. For example, the calculation apparatus 100 may register categories such as “person name”, “location”, and “money” in advance, and use information related to words corresponding to the category for processing.

[5-6. (Process by account)
In the said embodiment, it showed that the calculation apparatus 100 produced | generated a model based on the information regarding the person 60 who left the store. At this time, the calculation apparatus 100 may generate a model using information for each store account.

  For example, the store exit 60 may have a plurality of store accounts, but even if one of them is forcibly closed, other stores may have been operating normally. In this case, if all store accounts related to the store 60 are handled as store accounts that have been “forced to leave”, information about normal store accounts is also handled as store accounts that have been “forced to leave”, that is, correct data. The accuracy of the model may be reduced.

  Therefore, when the store exit 60 has a plurality of store accounts in a predetermined commercial transaction site, the calculation device 100 acquires information regarding the store exit 60 in the predetermined commercial transaction site for each store account. And the calculation apparatus 100 produces | generates a model by learning the account which satisfy | fills a predetermined condition as correct data among the store accounts which the shopkeeper 60 has. Specifically, the calculation device 100 is used for generating a model even if all the remaining store accounts are closed due to the forced closing of one store account of a predetermined store 60. As for information, only the store account that actually caused the forced exit is used. Thereby, the calculation apparatus 100 can generate a model that does not reduce the accuracy when obtaining the correlation with the store clerk 60.

[6. Others]
In addition, among the processes described in the above embodiment, all or part of the processes described as being automatically performed can be performed manually, or the processes described as being performed manually can be performed. All or a part can be automatically performed by a known method. In addition, the processing procedures, specific names, and information including various data and parameters shown in the document and drawings can be arbitrarily changed unless otherwise specified.

  Further, each component of each illustrated apparatus is functionally conceptual, and does not necessarily need to be physically configured as illustrated. In other words, the specific form of distribution / integration of each device is not limited to that shown in the figure, and all or a part thereof may be functionally or physically distributed or arbitrarily distributed in arbitrary units according to various loads or usage conditions. Can be integrated and configured.

  For example, the information in the storage unit 120 illustrated in FIG. 3 may be stored in an external storage server or the like without the calculation device 100. In this case, the calculation device 100 obtains various stored information by accessing the storage server.

  In addition, for example, the calculation device 100 described above includes a front-end server side that mainly executes exchanges with external devices such as acquiring various types of information and accepting calculation requests for the creditworthiness of businesses, You may distribute to the back end server side which performs a calculation process etc. In this case, for example, the front-end server includes at least the acquisition unit 131. Further, the back-end server has at least a calculation unit 134.

[7. Hardware configuration)
Further, the calculation apparatus 100 according to the embodiment described above is realized by a computer 1000 having a configuration as shown in FIG. 13, for example. FIG. 13 is a hardware configuration diagram illustrating an example of a computer 1000 that implements the functions of the calculation apparatus 100. The computer 1000 includes a CPU 1100, RAM 1200, ROM 1300, HDD 1400, communication interface (I / F) 1500, input / output interface (I / F) 1600, and media interface (I / F) 1700.

  The CPU 1100 operates based on a program stored in the ROM 1300 or the HDD 1400 and controls each unit. 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 a program executed by the CPU 1100, data used by the program, and the like. The communication interface 1500 receives data from other devices via the communication network 500 (corresponding to the network N shown in FIG. 2) and sends the data to the CPU 1100, and other data generated by the CPU 1100 via the communication network 500 To the device.

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

  The media interface 1700 reads a program or data stored in the recording medium 1800 and provides it to the CPU 1100 via the RAM 1200. The CPU 1100 loads such a program from the recording medium 1800 onto the RAM 1200 via the media interface 1700, and executes the loaded program. The recording medium 1800 is, for example, an optical recording medium such as a DVD (Digital Versatile Disc) or PD (Phase change rewritable disk), a magneto-optical recording medium such as 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 calculation device 100, the CPU 1100 of the computer 1000 implements the function of the control unit 130 by executing a program loaded on the RAM 1200. In addition, each data in the storage unit 120 is stored in the HDD 1400. The CPU 1100 of the computer 1000 reads these programs from the recording medium 1800 and executes them, but as another example, these programs may be acquired from other devices via the communication network 500.

[8. effect〕
As described above, the calculation device 100 according to the embodiment includes the acquisition unit 131 and the calculation unit 134. The acquisition unit 131 acquires information on the store exiter 60 who has lost the right to perform a commercial transaction at a predetermined commercial transaction site on the network. The calculation unit 134 calculates the credit rating of the store opening person 70 who opens a store on a predetermined commercial transaction site based on the information regarding the store closing person 60 acquired by the acquiring unit 131.

  As described above, the calculation device 100 according to the embodiment is based not only on financial information of a company that is generally used for measuring the creditworthiness of an operator, but also on information on the store exit 60 in a commercial transaction site such as a shopping mall. Thus, the creditworthiness of the shopkeeper 70 is calculated. In other words, the calculation device 100 assumes that the information about the business that has left the commercial transaction site is unfair information, and thus, based on the relevance with the information, determines the creditworthiness of the store 70 that opens the store on the commercial transaction site. calculate. For this reason, according to the calculation apparatus 100, normally, the creditworthiness with high accuracy can be provided even to the store opening person 70 who is a small business whose determination of creditworthiness is difficult unless it is based on reliable financial information or a predetermined white list. Can be calculated. Furthermore, the calculation device 100 has a configuration in which the credit rating is calculated based on the store clerk 60 at the commercial transaction site as a criterion, so that the credit rating can be calculated in accordance with the actual situation at the predetermined commercial transaction site.

  In addition, the calculation device 100 according to the embodiment is a model generated based on information about the store opening 60, and when information about the store opening 70 is input, the store opening 60 and the store opening 70 It further has a generating unit 132 that generates a model that outputs an index value indicating the correlation. The calculation unit 134 calculates the creditworthiness of the store seller 70 based on the index value (score) output from the model generated by the generation unit 132.

  As described above, the calculation device 100 according to the embodiment calculates the credit rating of the store seller 70 to be processed using the model generated based on the information regarding the store closing person 60. Thereby, the calculation apparatus 100 can calculate the creditworthiness of the storekeeper 70 efficiently, and obtains the correlation between the storekeeper 60 and the storekeeper 70 using the learned model, thereby calculating the calculation process. Can improve the accuracy.

  Moreover, the acquisition part 131 acquires the information regarding the evaluation from the user with respect to the shopkeeper 60 in a predetermined | prescribed commercial transaction site as information regarding the shopkeeper 60. FIG. The production | generation part 132 produces | generates a model based on the information regarding the evaluation from the user with respect to the shopkeeper 60 in a predetermined | prescribed commercial transaction site.

  As described above, the calculation device 100 according to the embodiment acquires a reaction from a user who uses a commercial transaction site via a network, and generates a model using the acquired information. Thereby, the calculation apparatus 100 can calculate the creditworthiness of the store seller 70 using data with high freshness such as a response from the user.

  Moreover, the acquisition part 131 is the information regarding the evaluation from the user with respect to the store 60, the number of the reactions from the user with respect to the store or product which the store 60 opened, or the store or product which the store 60 opened. The information about the evaluation value scored by the user is acquired.

  As described above, the calculation device 100 according to the embodiment acquires the number of reactions and scores from the user and uses them for the calculation process. Thereby, the calculation apparatus 100 can calculate the creditworthiness of the shopkeeper 70 using highly reliable data in which the reaction from the user is directly reflected.

  Further, the acquisition unit 131 changes the number of reactions from the user to the store or product opened by the store 60, or the evaluation value evaluated by the user for the store or product opened by the store 60 Get information about.

  As described above, the calculation apparatus 100 according to the embodiment performs calculation processing on information related to user evaluation using information that captures the transition of the information. In general, it is assumed that a business operator who leaves a commercial transaction site shows a tendency that the evaluation from the user gradually decreases or the product sales decrease. The calculation apparatus 100 uses the information that changes in this way as a feature to determine whether or not the store 70 will leave the store in the future, in other words, whether the store 70 is a reliable business operator. It is possible to accurately determine whether or not.

  Moreover, the acquisition part 131 acquires the word contained in the review from the user with respect to the store or the goods which the store closing person 60 opened as information regarding the evaluation from the user with respect to the store closing person 60. The generation unit 132 is based on the correlation between the word included in the review from the user for the store or product opened by the store 60 and the word included in the review from the user for the store or product opened by the store 70. To generate a model that outputs index values.

  As described above, the calculation apparatus 100 according to the embodiment generates a model using the word itself included in the review as a feature. For example, there is a possibility that a review content having a similar tendency is posted from a user to a business operator who leaves a business transaction site. The calculation device 100 can accurately calculate the creditworthiness of the shopkeeper 70 in the commercial transaction site that accepts the review by capturing the reaction indicated by the user with information called words.

  Moreover, the acquisition part 131 acquires the attribute information as a company of the storekeeper 60 in a predetermined | prescribed commercial transaction site as information regarding the storekeeper 60. FIG. The production | generation part 132 produces | generates a model based on the attribute information as a business person of the shopkeeper 60 in a predetermined | prescribed commercial transaction site.

  Thus, the calculation apparatus 100 according to the embodiment may generate a model using attribute information as a business operator. As a result, the calculation device 100 can generate a model reflecting information such as what kind of attributes an operator having a tendency to leave at a commercial transaction site to be processed. The trustworthiness reflecting the tendency (ease of leaving the store) indicated by the attribute information can be calculated for the store opening person 70.

  Further, the acquisition unit 131 acquires, as attribute information, the legal entity as the business operator of the store exit 60 or the type of business as the business operator of the store exit 60. The generation unit 132 determines whether the legality of the store 60 as the business and the business as the business of the store 70 or the business as the business of the store 60 and the business as the business of the store 70. A model that outputs an index value based on the correlation is generated.

  As described above, the calculation apparatus 100 according to the embodiment may generate a model reflecting the legal character or business type of the business operator. In general, since the legal entity and the type of business are highly likely to be an index indicating the trend of business operators on a commercial transaction site, the calculation apparatus 100 generates a model that takes such information into account, thereby improving the accuracy of the credit calculation process. Can be improved.

  Moreover, the acquisition part 131 acquires the information regarding the store exit 60 in a predetermined | prescribed commercial transaction site for every account, when the store exit 60 has a some account in a predetermined | prescribed commercial transaction site. The generation unit 132 generates a model by learning, as correct answer data, an account that satisfies a predetermined condition among accounts held by the store closing person 60.

  As described above, the calculation device 100 according to the embodiment corrects a store account that satisfies a predetermined condition (for example, a store account that has been forcibly closed) when the store 60 has a plurality of store accounts. The model may be generated by learning as data and not learning the other store account of the store 60 as correct data. As a result, the calculation device 100 can improve only the truly unsavory store account as the correct answer data, thereby improving the accuracy of the model.

  The embodiments of the present application and some of the modifications of the embodiments have been described in detail with reference to the drawings. The present invention can be implemented in other forms based on various modifications and improvements.

  The above-described calculation device 100 may be realized by a plurality of server computers, and depending on the function, the configuration may be realized by calling an external platform or the like with an API (Application Programming Interface) or network computing. Can be changed flexibly.

  Further, “section (module, unit)” described in the claims can be read as “means”, “circuit”, and the like. For example, the generation unit can be read as generation means or a generation circuit.

DESCRIPTION OF SYMBOLS 1 Calculation system 10 User terminal 30 Web server 100 Calculation apparatus 110 Communication part 120 Storage part 121 Store opening information storage part 122 Feature information storage part 123 Business information table 124 Store information table 125 Product information table 126 Word information table 127 Model storage part 128 Loan condition storage unit 130 Control unit 131 Acquisition unit 132 Generation unit 133 Reception unit 134 Calculation unit 135 Determination unit

Claims (11)

An acquisition unit for acquiring information on a store clerk who has lost the right to conduct a commercial transaction at a predetermined commercial transaction site on the network;
Based on the information about the store clerk acquired by the acquisition unit, a calculation unit that calculates a credit rating for the store clerk who opens a store on the predetermined commercial transaction site;
A calculation device comprising: A model that is generated based on information related to the store clerk, and that outputs an index value indicating a correlation between the store clerk and the store clerk when information related to the store clerk is input. Generating unit to generate,
Further comprising
The calculation unit includes:
Based on the index value output from the model generated by the generation unit, to calculate the creditworthiness of the store seller,
The calculation apparatus according to claim 1. The acquisition unit
As information about the store clerk, obtain information about the evaluation from the user for the store clerk in the predetermined commercial transaction site,
The generator is
Generating the model based on information from the user regarding the store exit in the predetermined commercial transaction site;
The calculation apparatus according to claim 2, wherein: The acquisition unit
As information on the evaluation from the user to the store clerk, the user scored the number of reactions from the user to the store or product opened by the store clerk or the store or product opened by the store clerk. Get information about evaluation values,
The calculation apparatus according to claim 3. The acquisition unit
To obtain information on the transition of the number of reactions from the user to the store or product opened by the store clerk, or the evaluation value transition of the user's score for the store or product opened by the store clerk,
The calculation apparatus according to claim 4. The acquisition unit
As the information about the evaluation from the user for the store clerk, the word included in the review from the user for the store or product opened by the store clerk,
The generator is
The index value is calculated based on a correlation between a word included in a review from a user for a store or a product opened by the store and a word included in a review from a user for the store or the product opened by the store. Generate a model to output,
The calculation device according to any one of claims 3 to 5, wherein The acquisition unit
As the information regarding the store exit, obtain attribute information as a business operator of the store exit in the predetermined commercial transaction site,
The generator is
Based on attribute information as a business operator of the store clerk in the predetermined commercial transaction site, the model is generated.
The calculation device according to any one of claims 2 to 6, wherein The acquisition unit
As the attribute information, the legal entity as the business of the store clerk, or the business type as the business of the store clerk,
The generator is
Based on the correlation between the legal entity as the business operator of the store exit and the legal personality as the business operator of the store or the business category as the business operator of the store exit and the business category as the business operator of the store owner Generating a model that outputs the index value;
The calculation apparatus according to claim 7. The acquisition unit
In the case where the store clerk has a plurality of accounts in the predetermined commercial transaction site, information on the store exit in the predetermined commercial transaction site is acquired for each account,
The generator is
The model is generated by learning, as correct data, an account that meets a predetermined condition among accounts that the store exit has,
The calculation device according to any one of claims 2 to 8. A calculation method executed by a computer,
An acquisition step of acquiring information about a store clerk who has lost the right to conduct a commercial transaction at a predetermined commercial transaction site on the network;
Based on the information about the store clerk acquired by the acquisition step, a calculation step of calculating a credit rating for the store entrepreneur who opens a store on the predetermined commercial transaction site;
The calculation method characterized by including. An acquisition procedure for obtaining information about a store losing the right to conduct business at a predetermined commercial transaction site on the network;
A calculation procedure for calculating a credit rating for a store clerk who opens a store on the predetermined commercial transaction site based on information on the store clerk acquired by the acquisition procedure;
A calculation program for causing a computer to execute.

Images