Credit support system, credit support method, and credit support program

Abstract

In order to enable prediction of the financial status of a company, a credit assistance system 100 is configured to include an extraction unit 101 and an output unit 102. The extraction unit 101 extracts a second company, financial indicators of which at a second point in time coincide with financial indicators of a first company at a first point in time later than the second point in time under prescribed conditions. The output unit 102 outputs actual data of the financial status of the second company after the second point in time.

Description

[Technical field]

[0001] The present invention relates to a technique for predicting the financial situation of a company. [Background technology]

[0002] When a financial institution conducts sales activities for financing a company, the sales activities are conducted on companies that are likely to close the deal depending on the economic situation and the financial situation of the company. In addition, when providing financing to a company, the screening department of the financial institution examines whether the company is suitable as a borrower, and provides credit if it is suitable. However, if the person in charge judges the economic and corporate situations and selects candidate companies, there is a risk that the candidate company selected may be a company that is not suitable for the situation or does not fit the overall strategy of the financial institution. In addition, when conducting sales activities related to financing, if the credit screening is not passed after preparations for financing are made, there is a risk that the efficiency of operations will decrease. Therefore, it is desirable to narrow down sales activities to companies that are likely to be accepted for financing and that are likely to be credited. As a technology for making predictions for such financing candidates, for example, a technology such as that in Patent Document 1 is disclosed.

[0003] The calculation device of Patent Document 1 calculates the degree of interest in financing from a second business operator as a degree of financing demand, based on the correlation between a first business operator that has provided financing in the past and a second business operator. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] JP 2018-116580 A Summary of the Invention [Problem to be solved by the invention]

[0005] However, although the technology of Patent Document 1 predicts the success or failure of a loan to a second business operator, it cannot predict the financial situation of the second business operator to which the loan is to be provided.

[0006] In order to solve the above problems, the present invention has an object to provide a credit support system and the like that can predict a company's financial situation from data on its financial indicators. [Means for solving the problem]

[0007] In order to solve the above problems, the credit support system of the present invention includes an extraction unit and an output unit. The extraction unit extracts a second company whose financial indicators at a first time point of a first company and at a second time point prior to the first time point match under a predetermined condition. The output unit outputs performance data on the financial status of the second company after the second time point.

[0008] The credit support method of the present invention extracts a second company whose financial indicators at a first time point of a first company and at a second time point prior to the first time point match under a predetermined condition. The credit support method of the present invention outputs performance data on the financial status of the second company after the second time point.

[0009] The program recording medium of the present invention records a credit support program. The credit support program causes a computer to execute a process of extracting a second company whose financial indicators at a first time point of a first company and at a second time point prior to the first time point match under a predetermined condition. The credit support program causes a computer to execute a process of outputting performance data on the financial status of the second company after the second time point. Effect of the Invention

[0010] According to the present invention, the financial situation of a company can be predicted from financial index data. [Brief description of the drawings]

[0011] [Figure 1] 1 is a diagram showing an outline of the configuration of a credit support system according to a first embodiment of the present invention. [Diagram 2] 1 is a diagram illustrating an example of the configuration of a credit support system according to a first embodiment of the present invention. [Diagram 3] FIG. 2 is a diagram illustrating an example of an operation flow of the credit support system according to the first embodiment of the present invention. [Figure 4] FIG. 2 is a diagram showing an example of a display screen in the first embodiment of the present invention. [Diagram 5] FIG. 2 is a diagram illustrating an example of an operation flow of the credit support system according to the first embodiment of the present invention. [Figure 6] FIG. 2 is a diagram showing an example of a display screen in the first embodiment of the present invention. [Figure 7] FIG. 2 is a diagram showing an example of a display screen in the first embodiment of the present invention. [Figure 8] FIG. 2 is a diagram showing an example of a display screen in the first embodiment of the present invention. [Figure 9] FIG. 2 is a diagram illustrating an example of an operation flow of the credit support system according to the first embodiment of the present invention. [Figure 10] FIG. 2 is a diagram illustrating an example of an operation flow of the credit support system according to the first embodiment of the present invention. [Figure 11] FIG. 11 is a diagram showing an outline of the configuration of a credit support system according to a second embodiment of the present invention. [Figure 12] FIG. 11 is a diagram illustrating an example of an operation flow of the credit support system according to the second embodiment of the present invention. [Figure 13] FIG. 13 is a diagram illustrating an example of another configuration of the embodiment of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0012] (First embodiment) A first embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing an outline of the configuration of a credit support system of this embodiment. The information processing system of this embodiment includes a credit support system 10 and a terminal device 20. The credit support system 10 and the terminal device 20 are connected via a network. Furthermore, the credit support system 10 is composed of one or more information processing devices. When composed of multiple information processing devices, the information processing devices are connected to each other via a network.

[0013] The information processing system of this embodiment is a system that predicts companies suitable as targets for loans and predicts the financial status of companies suitable as targets. The financial status refers to the financial growth or deterioration of a company. Financial growth means, for example, that when a figure in the business management increases, the figure of a desirable financial indicator also increases. Moreover, financial deterioration means, for example, that when a figure in the business management increases, the figure of a desirable financial indicator also decreases.

[0014] The credit support system 10 of the information processing system predicts companies suitable as loan recipients and their financial situations using a learning model generated by machine learning. The credit support system predicts candidates suitable as loan recipients by calculating a score indicating suitability as a loan recipient from data on the financial indicators of the companies using the learning model. In the following description, the learning model that calculates a score indicating suitability as a loan recipient from data on the financial indicators is referred to as a score calculation model. A candidate suitable as a loan recipient is, for example, a company that is likely to accept a loan when conducting sales activities and has the ability to repay. In addition, the credit support system uses the learning model to calculate the growth probability of each company from data on the financial indicators of the candidate companies and predicts their financial situations. In the following description, the learning model that calculates the growth probability of a company from data on the financial indicators and predicts its financial situation is referred to as a financial situation prediction model.

[0015] The following describes the configuration of the credit support system 10. Fig. 2 is a diagram showing an example of the configuration of the credit support system 10. The credit support system 10 includes an acquisition unit 11, a score calculation unit 12, a calculation model generation unit 13, a financial situation prediction unit 14, a prediction model generation unit 15, an extraction unit 16, an output unit 17, and a storage unit 18.

[0016] The acquisition unit 11 acquires, as input data, the selection result of the loan form input by the terminal device 20 through the operation of the operator and the selection result of the company for which the financial situation is predicted. The acquisition unit 11 also acquires data on the financial indicators of the company. Here, the financial indicator is a value that indicates the performance of the company during a period or the financial state at a certain point in time. The loan form refers to, for example, the method of loan and repayment according to the use of funds in the borrower company. The use of funds refers to, for example, an increase in working capital, repayment of other borrowings, raising funds for tax payment, or raising funds for acquisition. The use of funds is not limited to the above examples. The financial indicators to be referred to when examining candidates for loans and making credit decisions differ depending on the use of funds. Therefore, the credit support system 10 predicts companies suitable as targets for loans using a score calculation model set for each loan form.

[0017] The acquisition unit 11 acquires data indicating suitability as a loan target for a company using financial indicators as learning data in machine learning when generating a score calculation model predicting a company suitable as a loan target. The acquisition unit 11 acquires data indicating suitability as a loan target input to the terminal device 20 by an operator's operation from the terminal device 20. In addition, the acquisition unit 11 acquires data indicating the presence or absence of growth for a company using financial indicators as learning data in machine learning when generating a financial situation prediction model predicting the financial situation of a company. The acquisition unit 11 acquires data indicating the presence or absence of growth for a company using financial indicators as learning data in machine learning from the terminal device 20, which is input to the terminal device 20 by an operator's operation.

[0018] The score calculation unit 12 uses the financial indicator data of a company as input data and calculates a score indicating suitability as a loan target using a score calculation model. The score calculation model is generated in advance by machine learning for each loan type. The score calculation unit 12 extracts data of items set in the score calculation model from the financial indicator data as features. The score calculation unit 12 uses the extracted features as input data and calculates a score indicating suitability of each company as a loan target using the score calculation model.

[0019] The calculation model generation unit 13 performs machine learning using data on a company's financial indicators as input data and labels indicating whether the company is suitable as a loan target, and generates a score calculation model that predicts a score indicating the suitability of the company as a loan target from the data on the company's financial indicators.

[0020] The calculation model generation unit 13 classifies the input data of the financial indicators and the suitability as a loan target using a rule in the form of a decision tree as the correct answer data, and performs prediction using a linear model combining different explanatory variables in each case. The calculation model generation unit 13 generates a prediction model by sequentially performing the processes of optimizing the case classification conditions of the data, generating a prediction model by optimizing the combination of explanatory variables, and deleting unnecessary prediction models. This method of generating a prediction model using machine learning is also called heterogeneous mixture learning, because prediction is made by combining prediction models using combinations of different explanatory variables. The score calculation model may be generated by a machine learning algorithm other than the above method, as long as it can predict companies suitable as loan targets using the financial indicators of the companies. A heterogeneous mixture learning method is disclosed, for example, in the specification of U.S. Patent Application Publication No. 2014 / 0222741.

[0021] The calculation model generation unit 13 may re-learn the score calculation model using information on the success or failure of loans in actual sales activities. The calculation model generation unit 13 performs machine learning using data on the financial indicators of a company as input data and data indicating the suitability of a loan target reflecting actual performance as labels, and generates a score calculation model that calculates a score indicating suitability as a loan target from the data on the financial indicators of the company. The calculation model generation unit 13 uses the data of the score calculation model generated by re-learning to update the data of the score calculation model stored in the storage unit 18. By updating the score calculation model through re-learning using actual performance data, the accuracy of the estimation result of companies suitable as loan targets is further improved.

[0022] The financial situation prediction unit 14 uses the data of the financial indicators of the prediction target company as input data, and uses a financial situation prediction model to calculate the growth probability and deterioration probability after the present time, i.e., after the time indicated by the data of the financial indicators used in the prediction, to predict the financial situation of the target company. The financial situation prediction unit 14 may calculate only either the growth probability or the deterioration probability.

[0023] The financial situation prediction unit 14 calculates, for example, the growth probability and deterioration probability of the target company for each year from one year to three years after the year indicated by the data of the financial indicators used for the prediction. The prediction of the growth probability and deterioration probability does not have to be done on a yearly basis. In addition, the prediction period does not have to be three years.

[0024] The financial situation prediction unit 14 may predict the financial situation by scoring the growth probability and deterioration probability. The financial situation prediction unit 14 may also weight the rules used in calculating the growth probability and deterioration probability or the financial indicators included in the rules according to their empirical importance, and score the calculated growth probability and deterioration probability using a preset formula. The rules are composed of a combination of feature quantities used in the prediction and a condition that the combination of features satisfies.

[0025] When predicting the growth of a company, the financial situation prediction unit 14 uses data of the financial situation prediction model to extract features from data on financial indicators of the company to be predicted, and identifies rules in which the extracted features satisfy the set conditions. The financial situation prediction unit 14 selects a pre-set number of rules from the top of the rules whose features satisfy the conditions. The financial situation prediction unit 14 uses the growth probability and deterioration probability of each selected rule to calculate the average value of the growth probability and deterioration probability of all selected rules, and sets the calculated average value as the growth probability and deterioration probability of the company to be predicted. By predicting the growth probability and deterioration probability using a financial situation prediction model using such a machine learning algorithm, it is possible to predict the growth of a company while understanding the basis of the prediction using the rules.

[0026] The predictive model generation unit 15 uses data on the company's past financial indicators as input data and performs machine learning using actual data on growth or deterioration as labels to generate a predictive model that predicts the company's growth probability from its current financial indicators.

[0027] The prediction model generation unit 15 extracts data of preset items from data of financial indicators of a company as features. The prediction model generation unit 15 calculates a combination of features and a growth probability and a deterioration probability for each condition satisfied by the combination of features by a well-known method using random forest. The growth probability and the deterioration probability include a combination of features including a feature of only one financial indicator.

[0028] The prediction model generation unit 15 calculates a growth probability and a deterioration probability for each rule. The prediction model generation unit 15 calculates the growth probability using the ratio of the number of companies whose feature amounts satisfy the conditions set by the rules among the number of companies whose financial indicators were used in learning. The prediction model generation unit 15 calculates the deterioration probability using the ratio of the number of companies whose feature amounts do not satisfy the conditions set by the rules among the number of companies whose financial indicators were used in learning.

[0029] The prediction model generation unit 15 selects the generated rules using a selection algorithm, and identifies rules that are important in predicting the growth probability of the company. The prediction model generation unit 15 selects the rules using a selection algorithm that treats the rules as a continuous optimization problem, using, for example, a differentiable model using the probability calculated for each rule. The prediction model generation unit 15 ranks the selected rules, with the rules that have a higher degree of narrowing down being ranked higher. The prediction model generation unit 15 stores the rule rankings and data associated with each rule in the storage unit 18 as data for the financial situation prediction model. The machine learning algorithm generated in this way is also called rule discovery inference.

[0030] The prediction model generating unit 15 may generate a learning model that identifies categorical variables, which are items among financial indicators that have a high influence on the presence or absence of corporate growth, and supporting variables, which are items among other financial indicators that have a high influence on the categorical variables, and extracts variables that have a high influence on the prediction result. Here, the influence of a certain variable is the degree to which a change in the value of the variable changes the value of other variables. For example, a certain prediction model is generated by a machine learning algorithm and is an equation expressed as a weighted linear sum of explanatory variables. In that case, the influence of a certain explanatory variable is the weight coefficient of the explanatory variable, and is the degree to which a change in the value of the explanatory variable changes the value of the objective variable. In addition, when generating such a learning model, a pair of a categorical variable and a supporting variable may be configured in three layers by combining a supporting variable with a further variable as an auxiliary variable. The prediction model generating unit 15 performs prediction using a learning model generated by such a machine learning algorithm, and extracts a combination of explanatory variables that has a high influence on the presence or absence of corporate growth, which is the objective variable, thereby improving the explainability of the reason for the prediction result.

[0031] The extraction unit 16 extracts a company whose financial indicators at a certain time point in the past match those of the company to be predicted. For example, when the company to be predicted is a first company and the company whose financial indicators at a certain time point in the past match those of the first company is a second company, the extraction unit 16 extracts the second company whose financial indicators at a second time point prior to the first time point match those of the first company.

[0032] The extraction unit 16 uses, for example, a score indicating the financial status of each company calculated using a preset formula and the financial status of the prediction target company to identify companies whose current financial status and past financial status match. Financial indicators that match each other also include similar financial indicators that can be considered to match. For example, when the difference between the value obtained by scoring the financial indicator of the prediction target company and the value obtained by scoring the financial indicator at a certain point in time in the past is equal to or less than a preset standard, the extraction unit 16 extracts the company as one whose financial indicators match. That is, the extraction unit 16 extracts companies whose financial indicators of the prediction target company and financial indicators at a certain point in the past match within a predetermined condition.

[0033] The output unit 17 outputs information on the candidate companies according to the scores calculated by the score calculation unit 12. The output unit 17 ranks the candidate companies in descending order of score, and outputs a list of the company names arranged in rank order and the scores of each company as display data to the terminal device 20. The output unit 17 outputs, for example, display data of a list of company names and scores up to a specified rank to the terminal device 20. The output unit 17 may also output display data of a list of company names and scores for companies with scores equal to or higher than a standard.

[0034] The output unit 17 outputs data on the growth probability and deterioration probability of the target company. The output unit 17 also outputs reasons for the growth prediction and reasons for the deterioration prediction of the target company. The output unit 17 also outputs the transition of the financial status of companies whose past financial status matches the current financial status of the target company.

[0035] The storage unit 18 stores data of the score calculation model and the financial situation prediction model, and also stores data of financial indicators of the company.

[0036] The processes in the acquisition unit 11, score calculation unit 12, calculation model generation unit 13, financial situation prediction unit 14, prediction model generation unit 15, extraction unit 16, and output unit 17 can be performed, for example, by executing a computer program on a CPU (Central Processing Unit). Furthermore, the processes in the acquisition unit 11, score calculation unit 12, calculation model generation unit 13, financial situation prediction unit 14, prediction model generation unit 15, extraction unit 16, and output unit 17 may be performed in multiple information processing devices connected via a network.

[0037] When each process in the credit assistance system 10 is performed in a plurality of information processing devices, for example, the process in the score calculation unit 12 and the calculation model generation unit 13 and the process in the financial situation prediction unit 14, the prediction model generation unit 15, and the extraction unit 16 may be performed in different information processing devices. Also, for example, the process in the calculation model generation unit 13 and the prediction model generation unit 15 that generate a learning model by machine learning and the process in the score calculation unit 12, the financial situation prediction unit 14, and the extraction unit 16 that make predictions using the learning model may be performed in different information processing devices. In the above configuration, the acquisition unit 11, the output unit 17, and the storage unit 18 are provided in each information processing device. Also, the configuration when processing is performed in a plurality of information processing devices is not limited to the above example.

[0038] The storage unit 18 is configured using, for example, a hard disk drive. The storage unit 18 may be configured with other types of storage devices such as non-volatile semiconductor storage devices or a combination of multiple types of storage devices. The storage unit 18 may be provided on a storage device connected to the credit support system 10. The storage unit 18 may also be provided on a storage device controlled by an information processing device connected via a network.

[0039] The terminal device 20 displays on a display device (not shown) the list of candidate companies and the display data of the predicted financial status obtained from the credit assistance system 10. The terminal device 20 also sends the selection result of the loan type input by the operator's operation, data on the suitability of the company as a target for loan, data on the presence or absence of growth of the company, and the selection result of the company for which the financial status is predicted, as input data to the credit assistance system 10.

[0040] The operation of the information processing system of this embodiment will be described.

[0041] The operation of predicting suitability of a company as a loan target will be described below. Fig. 3 is a diagram showing an example of an operation flow when calculating a score indicating suitability as a loan target, among the operation flows of the credit support system 10.

[0042] The terminal device 20 receives the data of the selection result of the loan type input by the operator as input data. The terminal device 20 sends the input data of the selection result of the loan type to the credit processing system 10.

[0043] In FIG. 3, the acquisition unit 11 acquires data on the selection result of the loan type (step S11). The acquisition unit 11 also acquires data on financial indicators of a plurality of companies (step S12). The acquisition unit 11 acquires, for example, data on the financial indicators of the companies inputted to the terminal device 20 by the operation of an operator from the terminal device 20. The acquisition unit 11 may acquire data on the financial indicators of the companies from a server that provides financial information of the companies and is connected via a network. The data on the financial indicators of the companies may also be inputted to the credit support system 10 by the operation of an operator. The acquisition unit 11 stores the acquired data on the financial indicators in the storage unit 18. The companies from which the data on financial indicators is acquired may be set according to, for example, companies for which a financial institution may conduct sales activities for loans, or according to the type of business.

[0044] When the data of the loan type selection result is acquired, the score calculation unit 12 extracts data of items according to the settings of the selected loan type from the data of the company's financial indicators as feature quantities (step S13). When the data of items set by the loan type is extracted, the score calculation unit 12 uses the feature quantities extracted from the data of the financial indicators as input data and calculates a score indicating suitability as a loan target using a score calculation model (step S14). The score calculation unit 12 calculates the score indicating suitability as a loan target using the similarity of financial indicators with positive example companies.

[0045] When the score of a company is calculated, if there is a company whose score has not been calculated (No in step S15), the credit support system 10 repeats the process from the extraction of financial indicator data in step S12 onwards, and calculates the scores of the uncalculated companies.

[0046] When one score has been calculated, if the calculation of scores for all companies has been completed (Yes in step S15), the output unit 17 generates a list of candidate companies for loans, with company names and scores arranged in order of highest score, and outputs the list to the terminal device 20 (step S16).

[0047] When the list of candidate companies for loan recipients is received, the terminal device 20 displays the list of candidate companies for loan recipients on a display device (not shown).

[0048] FIG. 4 is a diagram showing an example of a display screen of the prediction result of the candidate companies to be loaned. On the left side of FIG. 4, the scenario name "Increase in Working Capital A" indicating the loan form and the rule applied when the score was calculated are displayed. On the right side of FIG. 4, the customer name indicating the name of the candidate companies and the score of each candidate company are displayed as a list in order of score. By displaying a display screen such as that of FIG. 4, the worker can obtain information on companies that are highly suitable as loan targets while referring to the rules at the time of prediction. In the example of FIG. 4, the applied rule may be configured to be rewritable by the operator's operation. When the applied rule is rewritable, the acquisition unit 11 acquires the change result of the rule input to the terminal device 20 by the operator's operation as input data. The calculation model generation unit 13 updates the data of the score calculation model using the acquired input data.

[0049] The operation of predicting the financial situation of a company will be described below. Fig. 5 is a diagram showing an example of the operation flow when predicting the financial situation of a company, among the operation flows of the credit support system.

[0050] The terminal device 20 acquires the selection result of the target companies for the financial situation prediction input by the operator as input data. The terminal device 20 sends the input data of the selection result of the target companies for the prediction to the credit support system 10. The target companies for the financial situation prediction are selected, for example, by the operator clicking the name of the candidate company with the mouse on the display screen of the list of candidate companies for loan recipients in Fig. 4.

[0051] In Fig. 5, the acquisition unit 11 acquires the selection result of the companies to be the targets of financial status prediction from the terminal device 20 (step S21). When the selection result of the companies to be predicted is acquired, the acquisition unit 11 acquires data on financial indicators of the selected companies (step S22). The acquisition unit 11 stores the acquired data on financial indicators of the companies in the storage unit 18. Furthermore, when the financial indicators of the selected companies are stored in the storage unit 18, the acquisition unit 11 does not need to acquire new data on financial indicators.

[0052] When the data of the financial indicators is acquired, the financial situation prediction unit 14 predicts the financial situation of the target company of the prediction using a financial situation prediction model with the data of the financial indicators as input data (step S23). The financial situation prediction unit 14 calculates the financial situation as a growth probability. When predicting the financial situation, the financial situation prediction unit 14 extracts a preset number of prediction rules in descending order of probability from among prediction rules whose conditions apply. The financial situation prediction unit 14 calculates the growth probability of the target company by calculating the average value of the growth probability for each extracted prediction rule. In addition, the financial situation prediction unit 14 calculates the deterioration probability of the target company by calculating the average value of the deterioration probability for each extracted prediction rule.

[0053] After calculating the growth probability and deterioration probability, the financial situation prediction unit 14 extracts the conditions indicated by the rules used in the calculation of the growth probability and deterioration probability as the reason for the prediction (step S24).

[0054] The extraction unit 16 also extracts companies whose current financial indicators of the prediction target company and their past financial indicators are highly similar as companies whose financial indicators match (step S25). The extraction unit 16 calculates the scores of the financial indicators of the companies whose financial indicators are highly similar using a preset calculation formula, for example, and extracts companies whose financial indicators are highly similar as companies whose financial indicators match using the calculated scores. The extraction unit 16 may specify companies that are included many times in the extracted prediction rules as companies whose financial indicators match those of the prediction target company.

[0055] When similar companies whose past financial situations match the predicted financial situation are extracted, the output unit 17 generates display data of the predicted financial situation results and outputs it to the terminal device 20. The output unit 17 outputs the predicted financial situation results of the prediction target company, performance data of companies whose financial indicators match, and the reasons for predicting the financial situation as prediction results to the terminal device (step S26). When the display data of the prediction results is acquired, the terminal device 20 uses the display data to display the prediction results on a display device (not shown).

[0056] FIG. 6 is a diagram showing an example of a display screen of the results of the forecast of financial status. In the upper left of FIG. 6, the names of the companies that are the targets of the forecast are displayed in the column of financial status forecast company. In the example of FIG. 6, company A is displayed as the target company of the forecast. In addition, under the name of "Financial target forecast company" in FIG. 6, the forecast results of the financial status of the target company of the forecast are displayed as growth probability and deterioration probability. In addition, on the right side of the upper part of FIG. 6, the names of companies with matching past financial indicators are displayed as a similar company list. The similar company list is composed of the names of companies with matching financial indicators, the years in which the financial indicators match, and a score indicating the similarity.

[0057] In the middle part of FIG. 6, the growth prediction application rule, which is the rule used when the growth probability is calculated, and the deterioration prediction application rule, which is the rule used when the deterioration probability is calculated, are displayed. The display field of the growth prediction application rule is composed of an ID indicating the identification of the rule, the content of the rule, and the probability of each rule. The rule with ID 10 includes a condition that the index of the cash / debt ratio is 90 or less, and the growth probability indicated by the rule with ID 10 is 80.5 percent. The display field of the deterioration prediction application rule is composed of an ID indicating the identification of the rule, the content of the rule, and the probability of each rule. The rule with ID 51 includes a condition that the debt dependency ratio is greater than 30.3, and the growth probability indicated by the rule with ID 51 is 1.5 percent. Note that only one of the growth prediction application rule or the deterioration prediction application rule may be displayed.

[0058] In the lower part of FIG. 6, reasons for growth prediction and deterioration prediction are shown as positive evaluation reasons and negative evaluation reasons. The positive evaluation reasons are output as natural sentences generated using a rule with a high probability among the growth prediction application rules. The correspondence between the items included in the rule and the natural sentences generated is set in advance. The correspondence between the items included in the rule and the natural sentences generated may be determined using a learning model generated by machine learning. The negative evaluation reasons are output as natural sentences generated using a rule with a high probability among the deterioration prediction application rules. The correspondence between the items included in the rule and the natural sentences generated is set in advance. Note that only one of the positive evaluation reasons or the negative evaluation reasons may be displayed.

[0059] Those involved in loan sales activities, or credit examiners who decide whether to grant loans to companies, can refer to prediction results like those in Figure 6 and use the prediction results to judge the validity of the prediction results using the reasons for the prediction, while also using the prediction results to judge the company's financial situation.

[0060] FIG. 7 is a diagram showing an example of a display screen of the prediction result when a similar company is selected in FIG. 6. In FIG. 7, the financial indicators of the prediction target company and the financial indicators of company G for fiscal year 2014 are highly similar, and company G is selected as a company with matching financial indicators. The selection of a company with matching financial indicators is performed, for example, by an operator clicking on the name of a similar company by operating the mouse on the display screen of the similar company list in FIG. 6. The selection of a company with matching financial indicators may also be performed by automatically selecting a company with the highest similarity in financial terms.

[0061] The upper graph in Figure 7 shows the growth probability of Company A predicted by the financial situation forecast model, scored using a specified formula, and the actual performance data of Company G, scored. The vertical axis in Figure 7 indicates the magnitude of the score. The base year for Company A is the current year, and for Company G, it is 2014, a year in which the financial indicators of Company A and Company G match. By displaying the years in which financial indicators match in this way, it is possible to compare the forecast results with the actual performance data and easily confirm the validity of the forecast results.

[0062] The left side of the lower part of FIG. 7 shows the financial status of Company A, the target company for the prediction, and the right side shows the financial status of Company G, which has similar financial indicators. The bar graphs in the graphs in the lower part of FIG. 7 show, for example, sales, and the line graphs show, for example, ordinary profit. The vertical axis of the graphs in the lower part of FIG. 7 is set to show, for example, amounts or standardized amounts. The output unit 17 displays the financial status of Company A in 2021, which is the current financial status, and the financial status of Company G in 2014, which matches the current financial status, so that the positions in the respective graphs match.

[0063] In the graph of Company A's growth forecast in FIG. 7, 2020 and 2021 are actual data, and 2022 and 2023 are forecast values. The future forecast of Company A's growth forecast is calculated, for example, using the statistical values ​​of the amount of change of similar companies. The future forecast of Company A's growth forecast may be performed by a forecasting method using a linear model. In addition, only past and present data of Company A may be displayed. As shown in FIG. 7, by displaying the current state and forecast value of the company to be forecasted and the actual data of companies with similar past financial indicators side by side, a worker who uses the forecast result can make a decision using the forecast result while confirming the validity of the forecast result.

[0064] FIG. 8 is a diagram showing an example of a display screen that displays the prediction reason displayed in FIG. 6 in more detail. In FIG. 8, a category, a supporting variable, and an auxiliary variable are displayed as a set as a reference index. The reference index displays the item name and value of each variable that has a large influence on the prediction result. The reason for the positive evaluation is displayed as an explanatory text generated using the data of the reference index. Moreover, the index evaluation is displayed as an index of the influence that the reference index has on the prediction result, calculated by a preset method.

[0065] The operation of generating a score calculation model will now be described with reference to Fig. 9, which is a diagram showing an example of the operation flow of the credit support system 10 when generating a score calculation model.

[0066] The terminal device 20 acquires data indicating suitability as a target for sales activities input by an operator as input data. The terminal device 20 sends the input data indicating suitability as a target for sales activities to the credit support system 10.

[0067] 9, the acquisition unit 11 acquires data on the financial indicators of a company and data indicating suitability as a loan target (step S31). The acquisition unit 11 acquires, for example, input data input to the terminal device 20 by an operator's operation from the terminal device 20. The acquisition unit 11 may also acquire data on the financial indicators of a company from a server of a financial information providing service via a network.

[0068] When the data of the financial indicators of the company is acquired, the calculation model generation unit 13 extracts data of items used for generating a score calculation model as features from the data of the financial indicators (step S32). The items of the financial indicators to be extracted as features are input to the terminal device 20 by the operation of a credit screening expert or an operator instructed by the expert, and are acquired from the terminal device 20 by the acquisition unit 11, for example.

[0069] After extracting the features used to generate the score calculation model, the calculation model generation unit 13 performs machine learning using the features extracted from the data on the company's financial indicators as input data and data labels of suitability as a target for sales activities for the company (step S33). The calculation model generation unit 13 generates a score calculation model that outputs a score indicating suitability as a target for sales activities from the data on the company's financial indicators through machine learning. The score calculation model is generated for each type of financing. The score calculation model is composed of items extracted as features from the financial indicators and data including score calculation rules. The score calculation model is generated by machine learning using, for example, a linear model. The score calculation model may also be generated by machine learning using a heterogeneous mixture learning method.

[0070] When the calculation model generation unit 13 generates the score calculation model, it verifies the accuracy of the score calculation model by using test data. When the accuracy of the generated score calculation model satisfies a preset standard (Yes in step S34), the calculation model generation unit 13 stores the data of the generated score calculation model in the storage unit 18 as data of a trained model (step S35).

[0071] When the accuracy of the generated score calculation model does not satisfy a preset standard (No in step S34), the acquisition unit 11 acquires change data of the feature items used for predicting the score (step S36). The acquisition unit 11 acquires, from the terminal device 20, change data of the feature items used for predicting the score, which is input to the terminal device 20 by an operator's operation.

[0072] Once the change data of the features has been acquired, the credit support system 10 returns to step 31 and repeats the operations subsequent to the process in which the acquisition unit 11 acquires data on the company's financial indicators and data indicating its suitability as a target for sales activities.

[0073] The operation of generating a financial situation prediction model will now be described with reference to Fig. 10, which is a diagram showing an example of the operation flow of the credit support system 10 when the credit support system 10 generates a financial situation prediction model.

[0074] The terminal device 20 acquires data indicating whether or not the company is growing, which is input by an operator, as input data. The terminal device 20 sends the input data indicating whether or not the company is growing to the credit assistance system 10.

[0075] In FIG. 10, the acquisition unit 11 acquires data on the financial indicators of the company and data indicating the presence or absence of growth (step S41). The acquisition unit 11 acquires data indicating the presence or absence of growth of the company from the terminal device 20. The acquisition unit 11 acquires data on the financial indicators of the company from the terminal device 20, which has been input to the terminal device 20 by an operator's operation, for example. The acquisition unit 11 may acquire data on the financial indicators of the company from a server that provides financial information of the company and is connected via a network. The data on the financial indicators of the company may also be input to the credit assistance system 10 by an operator's operation. The acquisition unit 11 stores the acquired data on the financial indicators in the memory unit 18.

[0076] When data on the company's financial indicators and data indicating the presence or absence of growth are acquired, the prediction model generation unit 15 extracts data of items set in advance from the data on the financial indicators as feature quantities (step S42). The items of the financial indicators to be extracted as feature quantities are input to the terminal device 20 by, for example, an expert in corporate finance or an operator instructed by the expert, and are acquired from the terminal device 20 by the acquisition unit 11.

[0077] After generating the features, the prediction model generation unit 15 executes machine learning using the features as input data and the data indicating the presence or absence of growth as labels (step S43). The prediction model generation unit 15 generates a financial situation prediction model that predicts the presence or absence of growth of a company from the feature data extracted from the financial indicators.

[0078] After generating the financial situation prediction model, the prediction model generation unit 15 verifies the accuracy of the financial situation prediction model using test data. If the accuracy of the generated financial situation prediction model satisfies a preset standard (Yes in step S44), the prediction model generation unit 15 stores the data of the generated financial situation prediction model in the storage unit 18 as data of the trained model (step S45).

[0079] When the accuracy of the generated financial situation prediction model does not satisfy a preset standard (No in step S44), the acquisition unit 11 acquires change data of the feature items used for predicting the financial situation (step S46). The acquisition unit 11 acquires from the terminal device 20 the change data of the feature items used for predicting the financial situation, which is input to the terminal device 20 by an operator.

[0080] When the change data of the features is acquired, the credit support system 10 returns to step 41 and repeats the operations subsequent to the process in which the acquisition unit 11 acquires data on the company's financial indicators and data indicating suitability as a loan target.

[0081] The credit support system of this embodiment uses a score calculation model generated for each loan type using the suitability of the loan recipient as a label to calculate a score indicating the suitability of the company as a loan recipient from the data of the company's financial indicators. In addition, the credit support system of this embodiment uses the calculated score to generate a list of companies that are highly suitable as loan recipients. By referring to the list of companies that are highly suitable as loan recipients and approaching companies with high scores, the possibility of loan acceptance increases and the possibility of rejection in credit screening decreases, thereby improving the efficiency of sales activities for loans.

[0082] The credit support system 10 of the information processing system of this embodiment predicts the growth probability of a company selected as a target company for prediction from the data of financial indicators of the target company, using a financial status prediction model generated using data of the company's financial indicators as input data and the presence or absence of growth as a label. In addition, the credit support system 10 of this embodiment outputs an explanation of the reason for the prediction, using financial indicators that have a large influence on the prediction result of the growth probability. By outputting the reason for the prediction together with the prediction result, the prediction result can be used while checking the validity of the prediction result.

[0083] The credit assistance system 10 outputs performance data on the financial status of a company that is the subject of a prediction and a company whose past financial status matches. By outputting the performance data of the companies whose financial status matches and the data of the company that is the subject of a prediction in a comparable state, the future financial status of the target company can be more easily recognized. Therefore, the credit assistance system 10 of the information processing system of this embodiment can predict the financial status of a company from the data on financial indicators of the company that is the subject of a prediction.

[0084] Second embodiment A second embodiment of the present invention will be described in detail with reference to the drawings. FIG. 11 is a diagram showing an example of the configuration of a credit support system 100 of this embodiment. The credit support system 100 of this embodiment includes an extraction unit 101 and an output unit 102. The extraction unit 101 extracts a second company whose financial indicators at a first time point of a first company and financial indicators at a second time point prior to the first time point match under a predetermined condition. The output unit 102 outputs performance data on the financial status of the second company after the second time point.

[0085] The extraction unit 16 in the first embodiment is an example of the extraction unit 101. Moreover, the extraction unit 101 is one aspect of the extraction means. The output unit 17 in the first embodiment is an example of the output unit 102. Moreover, the output unit 102 is one aspect of the output means.

[0086] The operation of the credit support system 100 will be described. Fig. 12 is a diagram showing an example of the operation flow of the credit support system 100. The extraction unit 101 extracts a second company whose financial indicators at a first time point of a first company and at a second time point prior to the first time point match under a predetermined condition (step S101). When the second company is extracted, the output unit 102 outputs performance data on the financial status of the second company after the second time point (step S102).

[0087] The credit assistance system 100 of this embodiment outputs the results of the situation of a second company whose financial indicators of the first company and financial indicators of a second starting point from the first time point in the past match under a predetermined condition. Therefore, by using the credit assistance system 100 of this embodiment, it is possible to predict the financial situation of the first company by referring to the past performance data of the financial indicators of the second company that are similar to the financial indicators of the first company.

[0088] Each process in the credit support system 10 of the first embodiment and the credit support system 100 of the second embodiment can be performed by executing a computer program on a computer. Fig. 13 shows an example of the configuration of a computer 200 that executes a computer program that performs each process in the credit support system 10 of the first embodiment and the credit support system 100 of the second embodiment. The computer 200 includes a CPU 201, a memory 202, a storage device 203, an input / output I / F (Interface) 204, and a communication I / F 205.

[0089] The CPU 201 reads out and executes a computer program for performing each process from the storage device 203. The CPU 201 may be configured by a combination of a CPU and a GPU (Graphics Processing Unit). The memory 202 is configured by a DRAM (Dynamic Random Access Memory) or the like, and temporarily stores the computer program executed by the CPU 201 and data being processed. The storage device 203 stores the computer program executed by the CPU 201. The storage device 203 is configured by, for example, a non-volatile semiconductor storage device. Other storage devices such as a hard disk drive may be used for the storage device 203. The input / output I / F 204 is an interface that receives input from an operator and outputs display data and the like. The communication I / F 205 is an interface that transmits and receives data to and from the terminal device 20. The terminal device 20 may also have a similar configuration.

[0090] The computer program used to execute each process may be stored in a recording medium and distributed. As the recording medium, for example, a magnetic disk such as a magnetic tape for recording data or a hard disk may be used. Also, an optical disk such as a CD-ROM (Compact Disc Read Only Memory) may be used as the recording medium. A non-volatile semiconductor memory device may also be used as the recording medium.

[0091] A part or all of the above-described embodiments can be described as, but is not limited to, the following supplementary notes.

[0092] [Appendix 1] An extraction means for extracting a second company whose financial indicators at a first time point of a first company and at a second time point prior to the first time point match under a predetermined condition; an output means for outputting performance data on the financial status of the second company after the second time point; A credit support system comprising:

[0093] [Appendix 2] The system further includes a prediction means for predicting the financial status of the first company after the first time point from the financial indicators of the first company at the first time point, using a prediction model that has learned the relationship between data on financial indicators of the first company and data indicating the growth or deterioration of the first company, The credit support system of claim 1, wherein the output means outputs the financial status of the first company after the first time point predicted by the prediction means and actual data on the financial status of the second company after the second time point.

[0094] [Appendix 3] 3. The credit support system according to claim 2, wherein the output means outputs a financial indicator that has a higher impact on the forecast result of the financial situation of the first company after the first point in time than other financial indicators.

[0095] [Appendix 4] The prediction means identifies a first financial indicator having a higher degree of influence on the prediction result than other financial indicators, and a second financial indicator having a higher degree of influence on the first financial indicator than other financial indicators, 4. The credit support system according to claim 3, wherein the output means outputs the first financial indicator and the second financial indicator.

[0096] [Appendix 5] The credit support system according to claim 4, wherein the output means outputs an explanatory text of the prediction result based on the first financial indicator and the second financial indicator identified by the prediction means.

[0097] [Appendix 6] 6. The credit support system according to any one of claims 2 to 5, further comprising a predictive model generation means for generating the predictive model by machine learning.

[0098] [Appendix 7] The present invention further includes a score calculation means for calculating a score indicating suitability as a loan target from the financial indicators of a company by using a score calculation model that has learned the relationship between data on the financial indicators of the company and data indicating suitability of the company as a loan target, The output means outputs a list of companies whose scores meet a predetermined criterion; 7. A credit support system as described in any one of Appendices 2 to 6, wherein the prediction means predicts the financial situation of a company included in the list as the first company from the first point in time onwards.

[0099] [Appendix 8] the output means outputs information on feature items of the score calculation model used in calculating the score and weights of the feature items; The credit support system according to claim 7, wherein the score calculation means calculates the score using the score calculation model that reflects the changed values ​​of the feature items and the weights of the feature amounts.

[0100] [Appendix 9] The score calculation means calculates the score using the score calculation model according to the loan form to the company, 9. The credit support system according to claim 7 or 8, wherein the output means outputs the list according to the loan type.

[0101] [Appendix 10] A credit support system as described in any one of Appendices 7 to 9, further comprising a calculation model generation means for generating the score calculation model, and updating the score calculation model by re-learning using data indicating the suitability of the company as a loan target, which is set based on whether or not the loan to the company is realized.

[0102] [Appendix 11] extracting a second company whose financial indicators at a first time point of a first company and at a second time point prior to the first time point match under a predetermined condition; Outputting performance data on the financial status of the second company after the second time point. Credit support methods.

[0103] [Appendix 12] predicting the financial status of the first company after the first time point from the financial indicators of the first company at the first time point using a prediction model that has learned the relationship between data on financial indicators of the first company and data indicating the growth or deterioration status of the first company; 12. The credit support method according to claim 11, which outputs a predicted financial status of the first company after the first time point and actual data on the financial status of the second company after the second time point.

[0104] [Appendix 13] 13. The credit support method according to claim 12, further comprising outputting a financial indicator that has a higher impact on the forecast result of the financial situation of the first company after the first time point than other financial indicators.

[0105] [Appendix 14] Identifying a first financial indicator having a higher impact on the forecast result than other financial indicators and a second financial indicator having a higher impact on the first financial indicator than other financial indicators; 14. The credit support method according to claim 13, further comprising outputting the first financial indicator and the second financial indicator.

[0106] [Appendix 15] The credit support method according to claim 14, further comprising outputting an explanation of the prediction result based on the identified first financial indicator and the identified second financial indicator.

[0107] [Appendix 16] Calculate a score indicating the suitability of the company as a loan target from the company's financial indicators using a score calculation model that has learned the relationship between data on the company's financial indicators and data indicating the suitability of the company as a loan target; outputting a list of companies whose scores meet a predetermined criterion; A credit support method as described in any one of Appendices 12 to 15, which predicts the financial situation of a company included in the list as the first company from the first point in time onwards.

[0108] [Appendix 17] outputting information on the feature items of the score calculation model used in calculating the score and the weights of the feature items; 17. The credit support method according to claim 16, wherein the score is calculated using the score calculation model that reflects the changed values ​​of the feature items and weights of the feature amounts.

[0109] [Appendix 18] Calculating the score using the score calculation model according to the type of financing given to the company; 18. The credit support method according to claim 16 or 17, wherein the list according to the loan type is output.

[0110] [Appendix 19] A credit support method described in any one of Appendices 16 to 18, which generates the score calculation model, and updates the score calculation model by re-learning using data indicating the suitability of the company as a loan target, which is set based on whether or not a loan is realized to the company.

[0111] [Appendix 20] A process of extracting a second company whose financial indicators at a first time point of a first company and at a second time point prior to the first time point match under a predetermined condition; outputting performance data on the financial status of the second company after the second time point; A program recording medium having recorded thereon a credit support program for causing a computer to execute the above.

[0112] The present invention has been described above by taking the above-mentioned embodiment as an exemplary example. However, the present invention is not limited to the above-mentioned embodiment. That is, the present invention can be applied in various aspects that can be understood by a person skilled in the art within the scope of the present invention. [Explanation of symbols]

[0113] 10 Credit Support System 11 Acquisition Department 12 Score calculation section 13 Calculation model generation unit 14 Financial Forecasting Department 15 Prediction model generation unit 16 Extraction part 17 Output section 18 Storage section 20 Terminal Equipment 100 Credit Support System 101 Extraction part 102 Output section 200 Computers 201 CPU 202 Memory 203 Storage device 204 Input / Output Interface 205 Communication I / F

Claims

1. A prediction means for predicting the financial situation of a first company after a first time point based on financial indicators of the first company at the first time point, and identifying a first financial indicator having a higher influence on the predicted result than other financial indicators, and a second financial indicator having a higher influence on the first financial indicator than the other financial indicators; an extraction means for extracting a second company whose financial indicators at the first time point of the first company and at a second time point prior to the first time point match under a predetermined condition; an output means for outputting the financial condition of the first company after the first time point predicted by the prediction means, performance data of the financial condition of the second company after the second time point, the first financial indicator, and the second financial indicator; A credit support system comprising:

2. The credit support system of claim 1, wherein the prediction means uses a prediction model that has learned the relationship between data on a company's financial indicators and data indicating the company's growth or deterioration status to predict the financial situation of the first company after the first point in time from the financial indicators of the first company at the first point in time.

3. The credit assistance system according to claim 2 , further comprising a predictive model generation means for generating the predictive model by machine learning.

4. 4. The credit support system according to claim 1, wherein the output means outputs an explanation of the prediction result based on the first financial indicator and the second financial indicator specified by the prediction means.

5. The present invention further includes a score calculation means for calculating a score indicating suitability as a loan target from the financial indicators of a company by using a score calculation model that has learned the relationship between data on the financial indicators of the company and data indicating suitability of the company as a loan target, The output means outputs a list of companies whose scores meet a predetermined criterion; 5. The credit support system according to claim 1, wherein the prediction means predicts the financial situation of a company included in the list as the first company from the first point in time onward.

6. the output means outputs information on feature items of the score calculation model used in calculating the score and weights of the feature items; 6. The credit support system according to claim 5, wherein the score calculation means calculates the score using the score calculation model that reflects changes to the feature items and weights of the feature amounts.

7. A computer comprising: A process of predicting a financial situation of a first company after a first time point based on financial indicators of the first company at the first time point, and identifying a first financial indicator having a higher influence on the prediction result than other financial indicators and a second financial indicator having a higher influence on the first financial indicator than other financial indicators; extracting a second company whose financial indicators at the first time point of the first company and at a second time point prior to the first time point match under a predetermined condition; Outputting the predicted financial status of the first company after the first time point, performance data of the financial status of the second company after the second time point, the first financial indicator, and the second financial indicator. Credit support methods.

8. A process for predicting the financial situation of a first company after a first time point from financial indicators of the first company at the first time point, and identifying a first financial indicator having a higher influence on the predicted result than other financial indicators, and a second financial indicator having a higher influence on the first financial indicator than the other financial indicators; A process of extracting a second company whose financial indicators at the first time point of the first company and financial indicators at a second time point prior to the first time point match under a predetermined condition; outputting the predicted financial status of the first company after the first time point, performance data of the financial status of the second company after the second time point, the first financial indicator, and the second financial indicator; A credit assistance program that causes a computer to execute the above.

Images