JP7360118B1 - Examination support device, examination support method, and examination support program - Google Patents

Abstract

[Problem] Detect client companies suspected of deteriorating financial conditions and support lending operations at financial institutions. [Solution] An examination operation support device in a financial institution according to the present invention includes an acquisition means for acquiring transaction data of liquidity transaction details of past window dressing companies and non-window dressing companies; A calculation means for calculating index data based on each transaction data, and a characteristic index that shows the data characteristics of the window dressing company based on the data difference between the window dressing company's index data and the non-window dressing company's index data. an acquisition means for acquiring transaction data of liquidity transaction details at a customer company; a calculation means for calculating index data of a characteristic index based on transaction data of the customer company; and an index of characteristic indicators at the customer company. It has a relation determining means for determining the relation between the data and the index data of the characteristic index of the destination company, and an output means for outputting the client company determined to be related by the relation determining means. [Selection diagram] Figure 5

Description

The present invention relates to an examination work support device, an examination work support method, and an examination work support program.

In corporate accounting, window dressing refers to, for example, manipulating numbers in financial statements to make a company's financial condition or business performance look better than it actually is, or to make it appear lower than it actually is, in order to give an advantage to bank loan decisions. This refers to intentionally distorting the calculation of profits on a company's financial statements.

As a technique related to this, for example, Patent Document 1 discloses that the characteristics of the data of the financial statements that are known to be fraudulent are found in advance, and it is determined whether the financial data of the financial statements to be diagnosed has the characteristics of the data. A financial analysis support system is described that automatically determines whether the

Furthermore, for example, Patent Document 2 discloses a data analysis device that provides a data analysis device to enable data obtained as a result of fraudulent settlement of accounts to be understood as an abnormal value by comparing financial data of multiple companies. The equipment is described.

For example, in Patent Document 3, in a fraudulent financial transaction detection program that detects fraudulent financial transactions (money laundering), there is information for acquiring difference pattern information regarding the transaction history in the bank account of a new detection target for which fraudulent transactions are newly detected. In the acquisition step, reference difference pattern information regarding the transaction history in the bank account of the person to be detected in the past acquired in the past and the degree of association of three or more levels with the possibility of fraudulent transactions of the person to be detected in the past are referred to, It is described that a computer is caused to execute a determination step of determining the possibility of fraudulent transactions by the new detection target based on the difference pattern information acquired in the information acquisition step.

Japanese Patent Application Publication No. 2002-312580 Japanese Patent Application Publication No. 2006-252259 Japanese Patent Application Publication No. 2021-144356

However, since the data used in the inventions described in Patent Documents 1 and 2 is financial data of annual financial statements for each business year, there is a problem that the timeliness of discovery of window dressing is lacking. It is also said that the current accounting system leaves room for companies to record profits as they see fit, and some point out that there is only so much that can be learned from financial statements regarding fraudulent accounting techniques.

On the other hand, since the data used in the invention described in Patent Document 3 is the transaction history in a bank account, it is more timely in detecting window dressing than financial data in annual financial statements. However, the invention described in Patent Document 3 has a different purpose in that it detects money laundering rather than corporate window dressing. Furthermore, in the invention described in Patent Document 3, the reference difference pattern information regarding the past transaction history in the bank account of the person to be detected that has been obtained in advance is the extracted difference in the transaction history between two or more terms. However, since fraudulent transactions are not necessarily included between two or more terms, in this case, it is not possible to detect fraudulent financial transactions (money laundering) even by referring to the reference difference pattern information. difficult.

Nowadays, in financial institutions, for example, those in charge of loan screening departments are required to perform efficient screening tasks, not only for accounting fraud but also for credit ratings (downgrades) due to deterioration of financial conditions. There is. The present invention has been proposed in view of the above points, and one aspect of the present invention is to detect client companies suspected of deteriorating financial conditions and to support financing operations at financial institutions.

In order to solve the above-mentioned problems, the screening work support device according to the present invention is a screening work support device in a financial institution, which acquires transaction data of liquidity transaction details of past window-dressing companies and non-windowing companies. an acquisition means, a calculation means for calculating index data based on the transaction data of each of the window dressing company and the non-window dressing company, and data of the index data of the window dressing company and the index data of the non-window dressing company. an identification means for identifying a characteristic index in which the data characteristics of the customer company are expressed based on the difference; an acquisition means for acquiring transaction data of the liquidity transaction details of the customer company; calculation means for calculating the index data of the characteristic index based on the characteristic index; a relationship determination means for determining the relationship between the index data of the characteristic index at the customer company and the index data of the characteristic index at the destination company; and the relationship determination means. and output means for outputting the customer companies determined to be related by the means.

According to an embodiment of the present invention, in one aspect, it is possible to detect a client company whose financial condition is suspected to be deteriorating, and to support financing operations at a financial institution.

FIG. 1 is a diagram illustrating a configuration example of an examination work support system according to the present embodiment. FIG. 2 is a diagram showing an example of the hardware configuration of an examination work support server according to the present embodiment. It is a diagram showing an example of the software configuration of the examination work support server according to the present embodiment. An example of customer information held by the financial institution system 10 according to the present embodiment is shown. An example of account information and transaction detail information possessed by the financial institution system 10 according to the present embodiment is shown. FIG. 3 is a diagram illustrating the entire process of detecting a potential decoration destination according to the present embodiment. FIG. 3 is a diagram showing calculation of feature indexes according to the present embodiment. FIG. 3 is a diagram illustrating a characteristic index identification process of a potential destination of window dressing by the examination work support server 20 according to the present embodiment. FIG. 3 is a diagram illustrating a process for detecting a potential destination of embellishment by the examination work support server 20 according to the present embodiment. FIG. 2 is a diagram illustrating a comparison of index data regarding transaction periodicity and transaction suddenness according to the present embodiment. FIG. 3 is a diagram showing a correlation matrix between indicators based on transaction data according to the present embodiment. It is a figure showing the difference of the correlation matrix of a non-decoration destination company and a decoration destination company concerning this embodiment. FIG. 3 is a diagram showing an example of a detection list of customer companies according to the present embodiment. FIG. 3 is a diagram showing an example of a detection list of customer companies according to the present embodiment. It is a diagram showing an example of the software configuration of the examination work support server according to the present embodiment. FIG. 2 is a diagram illustrating the entire process of detecting a destination whose credit rating is likely to be downgraded according to the present embodiment. FIG. 3 is a diagram illustrating characteristic index identification processing of a destination whose credit rating is likely to be downgraded by the examination work support server 20-2 according to the present embodiment. FIG. 7 is a diagram illustrating processing for detecting a credit rating downgrade destination by the examination work support server 20-2 according to the present embodiment.

Embodiments of the present invention will be described in detail with reference to the drawings.
[Embodiment 1]
<System configuration>
FIG. 1 is a diagram showing a configuration example of an examination work support system according to the present embodiment. The screening business support system 100 in FIG. 1 includes a financial institution system 10, a screening business support server 20, a processing DB 30, and a terminal 40, which are connected via a network 70.

The financial institution system 10 includes various systems and DBs (databases) owned by financial institutions such as banks. The financial institution system 10 includes, for example, a core system that manages various deposits such as ordinary deposits and time deposits, a sales management system that manages sales results of financial products, and a system that manages and proposes financial products according to the asset management needs of each customer. It has a front compliance system for managing negotiation history, a history system (CRM) for managing negotiation history, and each DB associated with each system. Further, each DB according to the present embodiment includes, for example, a DB of customer master data, transaction data, and financial data of a customer company in a financial institution. Financial data includes not only the client company's financial statements for each business year (balance sheet, income statement, cash flow statement, etc.), but also interim trial balance data that shows the progress of profits and losses during the year. Contains data regarding the customer company's finances.

The screening business support server (hereinafter also simply referred to as support server) 20 detects, based on the processed data in the processed DB 30, client companies that have accounts at financial institutions that are suspected of having fraudulently falsified financial data. It is a server device. In addition, the support server 20 creates a detection list (relevance/score) based on values (relevance/score) indicating signs/possibilities of financial data falsification on the terminal 40 of a person in charge of a financial institution (for example, a person in charge of a loan screening department). (generate) and display. Note that the window dressing according to the present embodiment is not limited to window dressing based on financial statements for each business year, but includes window dressing in general of financial data (that is, window dressing).

The processing DB 30 sequentially acquires raw data (called this way because it has not been processed) from each DB of the financial institution system 10 currently in operation by the financial institution, and converts the acquired raw data into a format that can be computationally processed by the support server 20. This is a DB that stores processed data. Machine learning methods often cannot handle raw input data such as text or time series data as is, so such unstructured data is converted into vectors through conversion processing. Although raw data and processed data may have different formats, the content and meaning of the data is the same.

The terminal 40 is, for example, a PC (personal computer), a smartphone, a tablet terminal, or the like, and is a user terminal used by, for example, a person in charge of a loan screening department. The person in charge accesses the support server 20 using the terminal 40 to obtain and display the detection list. As described above, the detection list is a list of client companies that have accounts at financial institutions that are predicted to have signs or possibilities of falsified settlements. By checking with customers on the detection list in order of priority, the person in charge can efficiently check for signs and possibilities of fraudulent settlements even within a limited time. In addition, by searching for a specific client company for which loan screening is to be performed, for example, the person in charge can check whether there are signs or possibilities of falsifying the company's financial results. A predetermined application program, a general-purpose web browser, or the like for acquiring and displaying the detection list from the support server 20 is installed in the terminal 40 in advance.

The network 70 is a communication network including wired and wireless communication networks. The network 70 includes, for example, the Internet, a public network, WiFi (registered trademark), and the like.

<Hardware configuration>
FIG. 2 is a diagram showing an example of the hardware configuration of the examination work support server according to this embodiment. As shown in FIG. 2, the examination work support server 20 includes a CPU (Central Processing Unit) 21, a ROM (Read Only Memory) 22, a RAM (Random Access Memory) 23, an HDD (Hard Disk Drive) 24, and a communication device. It has 25.

The CPU 21 executes various programs and performs arithmetic processing. The ROM 22 stores programs necessary for startup. The RAM 23 is a work area that temporarily stores processing performed by the CPU 21 and stores data. The HDD 24 includes RAID (Redundant Array of Independent Disks) and stores various data and programs. The communication device 25 communicates with other devices via the network 70.

<Software configuration>
FIG. 3 is a diagram showing an example of the software configuration of the examination work support server according to the present embodiment. The examination work support server 20 includes a transaction data acquisition section 201, an index data calculation section 202, a feature index identification section 203, a relation determination section 204, an output section 205, and a storage section 206 as main functional sections.

The transaction data acquisition unit 201 has a function of acquiring transaction data of liquidity transaction details of past window-dressing destination companies and non-windowing destination companies. Further, the transaction data acquisition unit 201 has a function of acquiring transaction data of liquidity transaction details of a client company.

The index data calculation unit 202 has a function of calculating index data based on transaction data of each of the window dressing destination company and the non-windowing destination company. The index data calculation unit 202 also has a function of calculating index data of characteristic indexes based on transaction data of client companies.

The characteristic index identification unit 203 has a function of identifying a characteristic index in which the data characteristics of the window-dressing destination company are expressed based on the data difference between the index data of the window-dressing destination company and the index data of the non-windowing destination company. There is.

The association determination unit 204 has a function of determining the association between the index data of the characteristic index in the customer company and the index data of the characteristic index in the destination company.

The output unit 205 has a function of outputting customer companies determined to be related by the relationship determination unit 204.

The storage unit 206 has a function of storing, in the HDD 24, information on the characteristic index of the destination company specified by the characteristic index specifying unit 203.

Note that each functional unit is realized by a computer program executed on hardware resources such as the CPU, ROM, and RAM of the computer that constitutes the support server 20. These functional units may be read as "means", "module", "unit", or "circuit". Further, each DB may be placed in the HDD 24 of the support server 20 or an external storage device on the network 70. Further, each functional unit of the support server 20 may not only be realized by a single server device, but also may be realized as a system consisting of a plurality of devices by distributing the functions.

(Data example of each DB)
An example of data items in the DB of the financial institution system 10 is shown. However, needless to say, the data items shown in FIG. 4 are merely examples, and other data items may be included.

FIG. 4A shows an example of customer information held by the financial institution system 10 according to the present embodiment. This is customer attribute information for each customer, and is registered mainly based on an application from the customer, such as when opening an account. Depending on the type of customer, there are individual customers and corporate customers, and in this embodiment, a corporate customer will be treated as an example, but it may also be an individual customer such as a sole proprietor.

FIG. 4B shows an example of account information and transaction detail information possessed by the financial institution system 10 according to this embodiment. For example, each account information held by each customer such as savings account, current account, time deposit account, foreign currency deposit account, credit card information, transaction details information including foreign exchange transactions (liquidity transaction details information), past This is information on the loans made. Not only the latest but also all past transaction details are recorded.

Here, for example, customer information is static and difficult to change, while among transaction data, liquidity transaction detail information such as the transaction details of each deposit account and credit card usage details is relatively dynamic and does not change easily. This is easy-to-understand information and is called "transaction data." Generally, transaction data is a type of data handled by corporate information systems, and refers to data that records details of events that occur in connection with business operations.

Transaction data according to the present embodiment refers to data that records details of events that occur in a financial institution when a customer company uses financial services, such as liquidity transaction statement information. Transaction data is data whose information changes (update frequency) relatively frequently on a daily basis, compared to information such as customer information, which does not change frequently from the time of registration. However, since the frequency of updates depends on the frequency of use of financial services by each customer, the specific frequency of updates does not matter.

The support server 20 sequentially acquires raw data from the DB of the financial institution system 10 being operated by the financial institution, and processes and converts the raw data acquired by the support server 20 into a format that can be calculated and processes the processed data into the processed DB 30. Update to. The update cycle can be set arbitrarily depending on the operation, such as every second, every few seconds to tens of seconds, every minute, every few minutes to tens of minutes, every hour, every few hours, every day, every predetermined day, etc. Although it is possible to set the update interval, the shorter the update interval, the better, and it is preferable that the update interval be updated in real time together with the update of the raw data on the DB of the financial institution system 10. As a result, the processed DB 30 is updated with data equivalent to the DB of the financial institution system 10 at each update interval, although there is a difference in format between raw data and processed data.

<Detection process for potential embellishments>
FIG. 5 is a diagram illustrating the entire process of detecting a potential embellishment destination according to this embodiment. FIG. 6 is a diagram showing calculation of feature indexes according to this embodiment. FIG. 7 is a diagram illustrating a characteristic index identification process of a potentially embellished destination by the examination work support server 20 according to the present embodiment.

S1: The support server 20 acquires transaction data of the financial institution's past window dressing companies and non-windowing companies as learning data (big data) from the processing DB 30 (originally transaction data DB). The transaction data according to the present embodiment is transaction data of transaction items whose information changes frequently (update frequency) on a daily basis in a financial institution, and can be obtained from liquidity transaction details information such as a savings account or a checking account. This is each transaction item data.

S2: The support server 20 calculates a feature amount (index feature amount) indicating the characteristics of a predetermined index based on the transaction data of the decorated company and the non-decorated company. For example, the feature values of indicators A, B, C, etc. are calculated based on the transaction data of the target company, and the characteristics of the same indicators A, B, C, etc. are calculated based on the transaction data of the non-deceptive target company. Calculate the amount.

For example, examples of index data based on transaction item data in a deposit account include the total amount of deposits, which is a total of deposit amounts, the total amount of withdrawals, which is a total of withdrawal amounts, the total amount of transfers, which is a total of transfer amounts, the number of deposits, which is a total of the number of deposits, and the total amount of withdrawals. This includes the number of withdrawals, which is calculated by counting the number of deposits, and the number of transfers, which is calculated by counting the number of transfers.

S3: The support server 20 identifies an index (hereinafter referred to as a characteristic index) with a large difference in index feature amount between the window dressing destination company and the non-windowing destination company. Characteristic indicators that show large differences between window dressing companies and non-window dressing companies are indicators that are prominently seen in window dressing companies as shown in transaction data analysis, and are indicators that are characteristic of window dressing companies. . Further, the number of feature indexes is not limited to one, and a plurality of feature indexes may be specified. Information on the characteristic index of the identified target company (including, for example, the index name and index characteristic amount (value)) is stored in the HDD 24.

The method of determining how large the difference is depends on each index, but if there is a difference (deviation) in the index feature values between the two companies from a certain standard value (predetermined value), the index is determined as having a large difference. It can be identified as a characteristic index. Further, for example, if the indicator feature range group at the window dressing destination company does not belong to the index feature range group at the non-windowing destination company at a certain rate or more, the index can be identified as a feature index with a large difference.

FIG. 8 is a diagram illustrating a process of detecting a potentially embellished destination by the examination work support server 20 according to the present embodiment.
S21: The support server 20 acquires transaction data (transaction data) of the customer company to be monitored from the processing DB 30 (originally transaction data DB). The transaction data according to this embodiment is, for example, each transaction item data that can be obtained from liquidity transaction detail information of a client company.

S22: The support server 20 calculates the feature amount of the characteristic index for the customer company identified in S3 based on the transaction data of the customer company. If there are a plurality of feature indicators, the support server 20 calculates the feature amounts of all the plurality of feature indicators specified in S3 based on the transaction data. A specific example of the feature index will be described later.

S23: The support server 20 calculates a score indicating the possibility of fabrication of the customer company based on the feature amount of the feature index of the transaction data of the customer company and the feature amount of the feature index of the transaction data of the destination company. do. The higher (closer) the degree of association (distance) between the feature amount of the feature index of the customer company and the feature amount of the feature index of the destination company is, the higher the score is calculated.

S24: The support server 20 arranges the customer companies in descending order of their scores and outputs them to the terminal 40. The person in charge uses the terminal 40 to view the output results, giving priority to customer companies with a predetermined score or higher or customer companies with a predetermined rank or higher that are suspected of being falsified, and refers to transaction data, financial data, etc. in more detail. You can check whether there is any embellishment by doing the following.

Note that the detection process (S21 to S24) for potential window-dressing destinations may be executed at the timing when a detection list acquisition request, which will be described later, is received from the terminal 40, or when transaction data on the DB of the financial institution system 10 is updated. It may also be executed each time the processing DB 20 is updated, or it may be executed periodically, for example, every day.

<Analysis and evaluation of transaction data differences>
In order to find characteristic indicators of transaction data between non-defining partner companies and non-defining partner companies, the applicant uses the liquidity transaction details information (transaction data) of customer companies held by financial institutions to identify the characteristics of transaction data between non-defining partner companies and We analyzed and evaluated the differences in transaction data between companies that have been window-dressing in the past.

(Feature index example 1)
In conventional qualitative analysis, the lack of planning for financial management is cited as one of the main factors in the evaluation results of companies that are window-dressing companies. More specifically, companies that have used falsification in the past are likely to have poor accounting management, poor financial management, frequent financing, lack of ability to carry out plans, and other reasons. It has become one. In the present embodiment, as a method for quantitatively analyzing the concept corresponding to the evaluation of lack of planning for fund management, we will analyze the time-series transactions of transaction data between non-fabricated target companies and window-dressed target companies. We calculated index data regarding customer companies' transaction periodicity and transaction suddenness, and compared and evaluated them.

FIG. 9 is a diagram illustrating a comparison of index data regarding transaction periodicity and transaction suddenness according to the present embodiment. Specifically, the table shown in Figure 9 shows indicators related to transaction relationships (e.g. remittance amount, number of remittances, credit payment amount, number of credit payments, deposit transfer The horizontal axis shows time in chronological order, and the vertical axis shows the variance of the indicators (indicators A, B, C, etc.). For each indicator, the larger the variance value, the more dispersion and drastic changes there are compared to past transactions.

As shown in Figure 9, when comparing non-decorated companies and predetermined companies, the overall trend is that non-decorated companies have smaller dispersion in each index, while decoy companies have smaller dispersion in each indicator. The dispersion is increasing. In other words, non-fabricated companies have cyclical transactions and are able to control their transactions in a planned manner, while companies that have non-fabricated transactions have a sudden nature and are often irregular, and are unable to control their transactions in a planned manner. It can be said that it is not. Specifically, the variance values of the indicators A, B, and D have a large difference between the non-decorated target companies and the window-decorated target companies, that is, the indicators A, B, and D can be said to be characteristic indicators.

In this way, characteristic indicators A, B, and D are indicators observed in the window dressing company that appeared in the data through analysis of transaction data between the window dressing company and the non-window dressing company. This is an indicator that can be called a characteristic. Therefore, among the transaction data of the customer company to be monitored, for example, characteristic indicators A, B, and D are calculated, and for the calculated characteristic indicators A, B, and D, the characteristic indicators A, B, and D of the client company in the past are By calculating the degree of association with the client company, it is possible to determine whether the customer company exhibits the same lack of planning for fund management seen in the client company, and by extension, the possibility of deception.

In S22, the support server 20 collects indicator data regarding transaction periodicity and transaction suddenness (for example, characteristic indicators A, B, and D) as the characteristic indicators specified in S3, based on the transaction data of the client company to be monitored. ) is calculated. In S23, the support server 20 calculates the variance value ( A score indicating the possibility of fabrication of the customer company is calculated by comparing the index feature values). In the case of the characteristic indicators A, B, and D, the higher the dispersion value of the characteristic index in the customer company (for example, the larger the dispersion of the index), the more likely the customer company is to be the target company, similar to the past target companies. A high score indicating a high degree of relevance is calculated.

(Feature index example 2)
Usually, there is a relationship between predetermined transaction items. For example, if the number of deposits into a regular account increases, the total amount of deposits will naturally increase, so there is a high correlation between them. On the other hand, it is intuitively difficult to say that the correlation between the number of deposits into a regular account, the number of withdrawals, the number of withdrawals and transfers, or the amount of withdrawals and transfers is either high or low. Specifically, by calculating the correlation values between the number of deposits and withdrawals from a regular account, the number of deposits and withdrawals from a regular account, and the number of deposits from a regular account and the amount of withdrawals and transfers, we can determine the level of correlation. becomes clear.

In other words, while there is a certain reasonable correlation between transaction items (indicators based on transaction items) from the standards of non-fabricated partner companies, in the case of window dressing companies, there is a correlation between these transaction items. Distortion may occur here and there. To make it easier to understand, in the case of a company that is a window dressing company, for example, although the number of deposits into ordinary accounts has increased, the amount deposited has not increased that much, and the originally high correlation between the relevant indicators is not seen (there is an abnormality between the relevant indicators). This is a case where a strong relationship can be seen.

FIG. 10 is a diagram showing a correlation matrix between indicators based on transaction data according to the present embodiment. Specifically, the table shown in Figure 10 shows each index (for example, number of deposits, number of deposits transferred, amount of deposits transferred, total amount deposited, number of withdrawals, This shows the correlation value (-1.0 to 1.0) between the aggregated indicators (indicators a, b, c, etc.). It is. The larger the correlation value, the higher the correlation between the two indicators. Conversely, the smaller the correlation value, the smaller the correlation between the two indicators, or the negative correlation value means a higher negative correlation.

FIG. 11 is a diagram showing the difference between the correlation matrices between the non-decorated target company and the decorated target company according to the present embodiment. An indicator with a large absolute value of the difference value is an indicator that was observed in the window dressing company based on the analysis of transaction data between the window dressing company and the non-windowing partner company, and is an indicator that can be said to be a characteristic of the window dressing company. It is.
Therefore, among the transaction data of the customer company to be monitored, for example, the characteristic index (x1, y1) is calculated, and the characteristic index ( By calculating the degree of association with .

In S22, the support server 20, based on the transaction data of the customer company to be monitored, generates index data (for example, characteristic indexes (x1, y1), etc.) regarding the relationship between the transaction data as the characteristic index specified in S3. ) is calculated. In S23, the support server 20 calculates the correlation value (index feature amount) of the index data based on the index data regarding the relationship between the transaction data at the customer company and the index data regarding the relationship between the transaction data at the destination company. A score indicating the possibility of fraud is calculated for the customer company. In the case of the characteristic index (x1, y1)..., the larger the absolute value of the correlation difference between the indicators (the less the client company has a certain reasonable degree of correlation) ), a high score is calculated indicating that the customer company has a high degree of relationship with the target company.

Note that the index data regarding transaction periodicity/transaction suddenness and the index data regarding the relationship between transaction data are merely examples of characteristic indexes. Through the processes of S1 to S3, the support server 20 identifies characteristic indicators of the garnished destination company that are expressed in data through transaction data analysis, including those that cannot be easily found by humans. In other words, the support server 20 uses so-called AI (for example, deep learning) to analyze transaction data (big data) between non-decorated destination companies and prefabricated destination companies, and determines characteristics of transaction data indicators that have large differences between non-decorated destination companies and disguised destination companies. The support server autonomously discovers and analyzes the amount, focuses on the characteristics of the discovered transaction data indicators, and calculates a score based on the degree of relevance to the customer company being monitored. However, on the other hand, if there is a characteristic index (known characteristic index) found in advance by a person, such as index data regarding transaction periodicity/transaction suddenness or index data regarding the relationship between transaction data, it will be identified in S3. It is also possible to input it into the support server 20 as a feature index.

Also, for convenience, in order to explain that the possibility of embellishment can be measured based on characteristic indicators, it is known in advance that indicator data regarding transaction periodicity and transaction suddenness, and index data regarding the relationship between transaction data are characteristic indicators. My explanation was based on this premise. However, in reality, through the processes of S1 to S3 executed by the support server 20, it is specified as a characteristic index seen in the destination company expressed in the form of data.

(Indicator example)
Examples of indicators that are aggregated based on liquidity transaction details information for savings accounts and checking accounts are shown below. Each index is index data that is aggregated and calculated from one or more transaction data. At customer companies, each transaction data (transaction data) changes as each company uses its account, and index data also changes.

・Deposit system Ordinary_Number of deposits, Ordinary_Deposit amount, Ordinary_Deposits_Number of transfers, Ordinary_Deposits_Transfer amount, Current account_Number of deposits, Current account_Deposit amount, Current account_Deposits_Number of transfers, Current account_Deposits_Transfer amount・Withdrawals Normal_Number of withdrawals, Normal_Withdrawal amount, Normal_Withdrawals_Number of transfers, Normal_Withdrawals_Transfer amount, Normal_Withdrawals_Number of funds transfers, Normal_Withdrawals_Funds transfer amount, Current account_number of withdrawals, current account_withdrawal amountCurrent account_withdrawals_number of transfers, current account_withdrawals_transfer amount, current account_withdrawals_number of fund transfers, current account_withdrawals_fund transfer amount <Scoring of client companies Output>
FIG. 12 is a diagram showing an example of a detection list of customer companies according to the present embodiment. For example, a person in charge of the loan screening department accesses the support server 20 using the terminal 40 and obtains and displays the detection list 500 on the screen. Furthermore, the person in charge can narrow down the search for client companies in the detection list by selectively inputting the narrowing down conditions 501 such as fiscal year end month and industry type, if necessary.

The detection list 500 shown in FIG. 12 targets client companies that have accounts at financial institutions, and detects signs and possibilities of falsified settlements among client companies that are about to close their accounts in March 2023. Predicted customer companies are listed in order of priority based on their scores. The client company detection list 500 includes items such as priority 500a, store number, customer number, corporate name, industry, score 500b, loan balance, deposit balance, and current bank ratio.

The priority level 500a indicates the priority order in which the person in charge should confirm the financial status (financial data) of the customer company. The score 500b is a score indicating the possibility of embellishment of the customer company calculated in S23. The higher (closer) the degree of association (distance) between the feature amount of the feature index of the customer company and the feature amount of the feature index of the destination company is, the higher the score is calculated. The person in charge can grasp the degree of possibility of window dressing for each client company based on specific numbers.

FIG. 13 is a diagram showing an example of a detection list of customer companies according to the present embodiment. By selecting one customer company from the detection list 500, the person in charge can display the financial status information 503 detected for the customer company.

In particular, the financial status information 503 shown in FIG. 13 includes information 503a regarding index characteristics. For example, in the client company, an "abnormal transaction" has been detected, and an "C" indicating a certain level of abnormal transaction and an alert message "Comparison index (characteristic index)" Normal_Withdrawal_Number of funds transfers. ”: Care must be taken as the characteristics are similar to the companies that have been used as decorations in the past. ” (the indicator data regarding transaction periodicity and transaction suddenness are similar to the characteristics shown by the companies with which the company has been embellished in the past) is displayed. For example, "D" indicating a certain level of abnormal transactions and an alert message "Comparison index (characteristic index) "Number of deposits and number of withdrawals from ordinary account": Because the characteristics are similar to the company with which the company has been manipulated in the past, Caution is required. ” (the indicator data regarding the relationship of the transaction data is similar to the characteristics shown by the company with which the company has been embellished in the past) is displayed. Then, "C" indicating the level of possibility of embellishment is displayed as "overall evaluation".

As described above, according to the detection list of client companies according to the present embodiment, the person in charge can check the customers on the detection list in order of priority, thereby efficiently detecting signs of falsified settlements even within a limited time. You can check whether it is possible. In addition, by searching for a specific client company for which loan screening is to be performed, for example, the person in charge can check whether there are signs or possibilities of falsifying the company's financial results.

As described above, the transaction data according to the present embodiment uses transaction data in a customer company's deposit account at a financial institution to perform company monitoring and calculate a score for measuring the possibility of fabrication. Transaction data cannot be tampered with by client companies, data is acquired and updated frequently, and the amount of data is large. This makes it possible to detect customer companies with the possibility of fraud, etc., in a timely manner and in accordance with the actual transaction status.

Although annual financial statements are easy to quantitatively evaluate from the perspective of traditional financial analysis, data can be falsified or falsely reported, and the frequency of data acquisition and updating is low, so customer companies may be susceptible to falsification. is difficult to detect in a timely manner.

In addition, the person in charge can focus on only those clients that exceed a certain scoring risk level from among a huge number of client companies. This makes it possible to carry out efficient screening operations and company monitoring, for example, by prioritizing screening from customers with high risks and being able to quickly notice the existence of risks.

(supplement)
As explained above, if a customer company determines that the feature amount of the feature index of transaction data at the customer company has a high degree of correlation with the feature amount of the feature index of transaction data at the customer company, the If the possibility is suspected, a score higher than a certain standard is calculated for the customer company (S23).

However, when the score is calculated again after that, it may not be as high as the previous score. A customer company whose score has decreased in this way can be said to have a decreased possibility of fabrication, and depending on the score, it can be said that the possibility of fabrication has been eliminated. The support server 20 maintains the score history calculated each time, and compares it with past scores. If the current score is significantly lower than the previous one (for example, than a certain standard value), the person in charge This can be explicitly displayed on the terminal 40. In addition, if a client company with a high score is flagged as a monitored company, if the current score drops significantly from the past score, the client company will be automatically excluded from the managed companies. You may also do so.

Furthermore, for the customer company whose current score has significantly decreased from its past score, the support server 20 determines that the feature amount of the characteristic index of the transaction data of the customer company is the same as the characteristic index of the transaction data of the "non-decoration destination company". The feature amount and the degree of association may be determined. It is a white judgment. If it is determined that the degree of relevance is high (if a high score is calculated), the client company does not have the same indicator characteristics as the target company, and is furthermore It is confirmed that characteristic indicators similar to those of window-dressing companies can be seen. As a result, the customer company can say with greater certainty that the possibility of falsification has decreased, and that the possibility of falsification has been eliminated depending on the score.

[Embodiment 2]
In this embodiment, the present invention is applied to credit rating examination at financial institutions. A credit rating (also referred to as a credit rating) is a standard used by banks, etc. to determine whether or not to provide a loan. For example, when a bank receives a loan application from a company, it needs to decide whether to lend to the company or not, and if so, how to set the interest rate. This is the rating for the company. Credit rating is a comprehensive analysis of a company's solvency and financial situation, and is mainly an evaluation based on the analysis results of financial statements (quantitative evaluation) and an evaluation of management's attitude and management policy (qualitative evaluation). It is determined by two types of evaluation.

However, in evaluations (quantitative evaluations) based on analysis results of financial statements, it is difficult to make appropriate ratings when there is falsification or false reporting such as falsification of financial statements, and the frequency of acquiring and updating financial statements is low (business (every fiscal year), it is difficult to timely rate and evaluate companies. Therefore, the screening work support system 100-2 (screening work support server 20-2) according to the second embodiment supports credit rating screening by, for example, a person in charge of a loan screening department.

The standards for credit ratings vary depending on the bank or company doing the rating, but they are generally expressed in alphabetical or numerical ranks based on creditworthiness. In this embodiment, for convenience of explanation, it is assumed that there are six rating ranks, A to E, as an example. However, in reality, in many cases, a company's credibility is further subdivided into around 10 rating ranks.
A: (Normal) Financial conditions are good and the ability to pay is high. B: (Requires caution) Companies with poor or unstable business performance or problems with financial conditions, requiring careful management in the future. C: (Bankruptcies at risk) Not currently in a state of bankruptcy, but there is a high possibility that their ability to pay will decline in the future D: (Virtually bankrupt) There are problems with their ability to pay at present E: (Bankrupt) Already In this embodiment, a company with a credit rating of A is called a company with a normal credit rating (hereinafter simply referred to as a normal company), and a company with a credit rating other than A is called a company with an abnormal credit rating. This is called a company (hereinafter also simply referred to as an abnormal company).

<Software configuration>
FIG. 14 is a diagram showing an example of the software configuration of the examination work support server according to this embodiment. The examination work support server 20-2 has, as main functional units, a transaction data acquisition unit 201-2, an index data calculation unit 202-2, a characteristic index identification unit 203-2, a relation determination unit 204-2, and an output unit 205-. 2, and a storage section 206-2.

The transaction data acquisition unit 201-2 has a function of acquiring transaction data of liquidity transaction details of abnormal and normal credit rating companies in the past. Further, the transaction data acquisition unit 201-2 has a function of acquiring transaction data of liquidity transaction details of the client company.

The index data calculation unit 202-2 has a function of calculating index data based on the transaction data of each of the abnormal client company and the normal client company. In addition, the index data calculation unit 202-2 has a function of calculating index data of characteristic indexes based on transaction data of client companies.

The characteristic index identification unit 203-2 has a function of identifying a characteristic index in which the data characteristics of the abnormal client company are expressed based on the data difference between the index data of the abnormal client company and the index data of the normal client company. have.

The association determination unit 204-2 has a function of determining the association between the index data of the characteristic index in the client company and the index data of the characteristic index in the abnormal client company.

The output unit 205-2 has a function of outputting customer companies determined to be related by the relationship determination unit 204.

The storage unit 206-2 has a function of storing, in the HDD 24, information on the characteristic index of the abnormal company identified by the characteristic index identification unit 203.

Note that each functional unit is realized by a computer program executed on hardware resources such as the CPU, ROM, and RAM of the computer that constitutes the support server 20-2. These functional units may be read as "means", "module", "unit", or "circuit". Further, each DB may be placed in the HDD 24 of the support server 20-2 or an external storage device on the network 70. Further, each functional unit of the support server 20-2 is not only realized by a single server device, but also may be realized as a system consisting of a plurality of devices by distributing the functions.

Furthermore, in the second embodiment, the configuration example of the examination work support system shown in FIG. 1 and the hardware configuration example of the examination work support server shown in FIG. Omitted. However, the examination work support server 20 is referred to as an examination work support server 20-2 in the second embodiment.

<Detection process for potential credit rating downgrades>
FIG. 15 is a diagram showing the entire process of detecting a destination whose credit rating is likely to be downgraded according to this embodiment. FIG. 16 is a diagram illustrating a characteristic index identification process of a destination whose credit rating is likely to be downgraded, performed by the examination work support server 20-2 according to the present embodiment. Here, the companies that should be detected as soon as possible by the person in charge during the screening process are customer companies that are likely to be downgraded to a lower credit rating than their current credit rating. These are companies that used to be rated A or B, but have recently suddenly fallen into a financial position equivalent to a credit rating of B or C or lower.

S31: The support server 20-2 obtains transaction data of past companies with abnormal credit ratings and companies with normal credit ratings at financial institutions as learning data (big data) from the processing DB 30 (originally transaction data DB). get. Note that a company with an abnormal credit rating is, for example, a company with a credit rating of B, C, D, and/or E, and a normal company is, for example, a company with a credit rating of A.

S32: The support server 20-2 calculates a feature amount (index feature amount) indicating the characteristics of a predetermined index based on the transaction data of the abnormal and normal credit rating companies. For example, the feature quantities of indicators A, B, C, etc. are calculated based on the transaction data of a company with an abnormal credit rating, and the same indicators A, B, etc. are calculated based on the transaction data of a company with a normal credit rating. , C... is calculated.

S33: The support server 20-2 identifies characteristic indicators with large differences in index feature quantities between the non-normal credit rating company and the normal credit rating company. Characteristic indicators that have a large difference between abnormal and normal client companies are indicators that are conspicuously seen in the abnormal client companies as expressed in data through transaction data analysis, and are indicators that are characteristic of the abnormal client companies. It is. Further, the number of feature indexes is not limited to one, and a plurality of feature indexes may be specified. Information on the characteristic index of the identified abnormal company (including, for example, the index name and index characteristic amount (value)) is stored in the HDD 24.

- Note that in S31, if the company with an abnormal credit rating is, for example, a company with a credit rating of C, the normal company may be a company with a credit rating of A or B, for example. In other words, if a company ranks higher than abnormal companies, it is relatively normal.

-Furthermore, characteristic indicators that have a large difference between abnormal client companies and normal client companies may be relatively specified for each credit rating. For example, a characteristic index that has a large difference between an abnormal client company with a credit rating of B and a normal client company with a credit rating A, a characteristic index that has a large difference between an abnormal client company with a credit rating C and a normal client company with a credit rating A, or a credit rating. For example, there are characteristic indicators that have a large difference between an abnormal client company with a credit rating of C and a relatively normal client company with a credit rating of B. In this case, the information on the characteristic index of the specified abnormal client company is stored in the HDD 24 as a data set with both the credit rating information of the abnormal credit rating company and the normal client company.

・However, for example, a characteristic index that has a large difference between an abnormal customer company with a credit rating of D and a normal customer company with a credit rating A, or a characteristic index that has a large difference between an abnormal customer company with a credit rating E and a normal customer company with a credit rating A. However, the financial conditions of the two companies are clearly very different in the first place. Therefore, the closer the credit ratings between companies, the better from the point of view of identifying characteristic indicators of abnormal partner companies that are expressed in data through transaction data analysis, including those that cannot be easily found by humans. (For example, an abnormal client company with a credit rating of B, a normal client company with a credit rating of A, an abnormal client company with a credit rating of C, and a normal client company with a credit rating of A). Furthermore, characteristic indicators that have a large difference between companies whose credit ratings differ by one rank (companies whose credit ratings are adjacent by one rank) may be identified (for example, an abnormal company with a credit rating of B and an abnormal company with a credit rating of B). normal client companies with credit rating A, abnormal client companies with credit rating C, and normal client companies with credit rating B).

FIG. 17 is a diagram showing a process for detecting a credit rating downgrade destination by the examination work support server 20-2 according to the present embodiment.

S41: The support server 20-2 acquires the credit rating (FIG. 4B) of the customer company to be monitored from the customer master DB. The credit rating according to this embodiment indicates one of A to E.

S42: Next, the support server 20-2 acquires transaction data (transaction data) of the customer company to be monitored from the processing DB 30 (originally transaction data DB). The transaction data according to this embodiment is, for example, each transaction item data that can be obtained from liquidity transaction detail information of a client company.

S43: The support server 20-2 calculates the feature amount of the characteristic index for the customer company specified in S33, based on the transaction data of the customer company. If there are a plurality of feature indicators, the support server 20-2 calculates the feature amounts of all the plurality of feature indicators specified in S33 based on the transaction data.

S44: The support server 20-2 determines the possibility of a credit rating downgrade of the customer company based on the feature amount of the feature index of the transaction data of the customer company and the feature amount of the feature index of the transaction data of the abnormal client company. Calculate the score shown. The higher (closer) the degree of association (distance) between the feature quantity of the characteristic index of the customer company and the feature quantity of the characteristic index of the abnormal client company is, the higher the score is calculated.

Note that if the customer company has a credit rating of A, the abnormal client company may have a credit rating of B or lower than B. In this case, the support server 20-2 provides a score indicating the possibility of a downgrading of the customer company's credit rating based on the characteristic index of the transaction data of the customer company and the characteristic index of the transaction data of companies with credit ratings of B or lower. Calculate.

However, if the credit rating of the customer company acquired in S41 is, for example, B, the abnormal company from the perspective of the customer company is a company with a credit rating of C or lower. In this case, the support server 20-2 provides a score indicating the possibility of downgrading of the customer company's credit rating based on the characteristic index of the transaction data of the customer company and the characteristic index of the transaction data of companies with credit ratings of C or lower. can be calculated.

S45: The support server 20-2 arranges the customer companies in descending order of their scores and outputs them to the terminal 40. The person in charge uses the terminal 40 to view the output results, and gives priority to customer companies with a predetermined score or higher or customer companies with a predetermined rank or higher that are likely to have their credit ratings downgraded, and refers to transaction data, financial data, etc. in more detail. You can check your credit rating.

Note that the processing for detecting a party whose credit rating may be downgraded (S41 to S45) may be executed at the timing when a request to obtain a detection list of a party whose credit rating is likely to be downgraded is received from the terminal 40, or may be executed at the timing when a request to obtain a detection list of a party whose credit rating is likely to be downgraded is received, or when a transaction on the DB of the financial institution system 10 is performed. It may be executed each time the processing DB 20 is updated with data update, or it may be executed periodically, for example, every day.

As described above, the transaction data according to the present embodiment uses transaction data in a customer company's deposit account at a financial institution to perform company monitoring and calculate a score for measuring the possibility of credit rating downgrade. Transaction data cannot be tampered with by client companies, data is acquired and updated frequently, and the amount of data is large. This makes it possible to timely detect client companies that are at risk of having their credit ratings downgraded, etc., in line with the actual transaction status.

Although annual financial statements are easy to quantitatively evaluate from the perspective of conventional financial analysis, data can be falsified or falsely reported, and data is not acquired or updated frequently, so customers are at risk of having their credit ratings downgraded. It is difficult to detect companies in a timely manner.

In addition, the person in charge can focus on only those clients that exceed a certain scoring risk level from among a huge number of client companies. This makes it possible to carry out efficient screening operations and company monitoring, for example, by prioritizing screening from customers with high risks and being able to quickly notice the existence of risks.

(Supplement 1)
As explained above, in a client company, if it is determined that the feature amount of the feature index of the transaction data at the customer company is highly related to the feature amount of the feature index of the transaction data at the abnormal partner company, A score higher than a certain standard is calculated for the customer company as a case where the possibility of credit rating downgrade is suspected (S43).

However, when the score is calculated again after that, it may not be as high as the previous score. Client companies whose scores have fallen in this way can be said to have had their credit ratings lowered, or, depending on their scores, to have a possibility of having their credit ratings upgraded. The support server 20 maintains the score history calculated each time, and compares it with past scores. If the current score is significantly lower than the previous one (for example, than a certain standard value), the person in charge This can be explicitly displayed on the terminal 40. In addition, if a client company with a high score is flagged as a monitored company, if the current score drops significantly from the past score, the client company will be automatically excluded from the managed companies. You may also do so.

Further, for the customer company whose current score has significantly decreased from its past score, the support server 20-2 determines that the feature amount of the feature index of the transaction data of the customer company is a "normal customer company" (for example, a credit rating of A). The feature amount and degree of association of the feature index of transaction data in a company) may be determined. It is a white judgment. If it is determined that the degree of relevance is high (if a high score is calculated), the customer company will not have the characteristic indicators of the abnormal partner company, and will also It is confirmed that characteristic indicators similar to those of normal partner companies can be seen. As a result, the client company can say with greater certainty that the possibility of a credit rating downgrade has decreased, and that depending on the score, there is a possibility of a credit rating upgrade.

(Supplement 2)
As a modification of S44, when the credit rating of the customer company is an abnormal company (for example, credit rating of B or lower), the support server 20-2 calculates the feature amount of the characteristic index of the transaction data of the customer company and the credit rating. Calculate a score indicating the "possibility of credit rating upgrade" of the customer company based on the characteristic amount of the characteristic index of the transaction data of the "normal company" (for example, credit rating A) that is higher than the customer company. is also possible. The higher (closer) the degree of association (distance) between the feature quantity of the characteristic index of the customer company and the feature quantity of the characteristic index of the normal client company is, the higher the score is calculated. The support server 20-2 arranges the customer companies in descending order of their scores and outputs them to the terminal 40.

As a result, the person in charge uses the terminal 40 to view the output results, and prioritizes customer companies with a predetermined score or higher or customer companies with a predetermined rank or higher that are considered likely to have a credit rating upgraded, and provides more detailed transaction data, financial data, etc. You can review and confirm your credit rating by referring to . In addition, the person in charge can focus on only those clients that fall below a certain scoring risk level from among a huge number of client companies. For example, it is possible to carry out efficient screening operations and company monitoring, such as giving priority to screening customers whose risks have decreased and quickly noticing reductions in risk.

<Summary>
As described above, according to the screening business support system according to the present embodiment, it is possible to detect a client company whose financial condition is suspected to be deteriorating, and to support financing operations at a financial institution. Although the present invention has been described with reference to specific examples according to preferred embodiments of the present invention, it is to be understood that these specific examples may be modified without departing from the broad spirit and scope of the present invention as defined in the claims. Obviously, various modifications and changes may be made to the examples. In other words, the invention is not to be construed as limited by the details of the specific examples and the accompanying drawings.

<About the invention group extracted from this embodiment>
The characteristics of the invention group extracted from the above-described embodiments will be described below.
(Appendix A1)
An examination support device for financial institutions,
an acquisition means for acquiring transaction data of liquidity transaction details of past window dressing companies and non-windowing partner companies;
Calculation means for calculating index data based on the transaction data of each of the window dressing company and the non-windowing company;
Identification means for identifying a characteristic index in which the data characteristics of the window dressing company are expressed based on a data difference between the index data of the window dressing company and the index data of the non-windowing window company;
an acquisition means for acquiring transaction data of liquidity transaction details at a customer company;
Calculation means for calculating index data of the characteristic index based on the transaction data of the customer company;
association determining means for determining a relationship between index data of the characteristic index at the customer company and index data of the characteristic index at the destination company;
Output means for outputting the customer companies determined to be related by the relationship determination means;
An examination work support device characterized by having the following.

(Appendix B1)
An examination support device for financial institutions,
an acquisition means for acquiring transaction data of liquidity transaction details of past non-normal credit rating companies and normal partner companies;
Calculation means for calculating index data based on the transaction data of each of the abnormal partner company and the normal partner company;
Identification means for identifying a characteristic index in which the data characteristics of the abnormal client company are expressed based on a data difference between the index data of the abnormal client company and the index data of the normal client company;
an acquisition means for acquiring transaction data of liquidity transaction details at a customer company;
Calculation means for calculating index data of the characteristic index based on the transaction data of the customer company;
association determining means for determining a relationship between index data of the characteristic index at the customer company and index data of the characteristic index at the abnormal client company;
Output means for outputting the customer companies determined to be related by the relationship determination means;
An examination work support device characterized by having the following.

10 Financial institution system 20, 20-2 Screening operations support server 30 Processing DB
40 Terminal 70 Network 100, 100-2 Screening work support system 201, 201-2 Transaction data acquisition section 202, 202-2 Index data calculation section 203, 203-2 Feature index identification section 204, 204-2 Association determination section 205, 205-2 Output section 206, 206-2 Storage section

Claims (7)

An examination support device for financial institutions,
an acquisition means for acquiring transaction data of liquidity transaction details of past window dressing companies and non-windowing partner companies;
Calculation means for calculating index data based on the transaction data of each of the window dressing company and the non-windowing company;
Identification means for identifying a characteristic index in which the data characteristics of the window dressing company are expressed based on a data difference between the index data of the window dressing company and the index data of the non-windowing window company;
an acquisition means for acquiring transaction data of liquidity transaction details at a customer company;
Calculation means for calculating index data of the characteristic index based on the transaction data of the customer company;
a relationship determining means for determining a relationship between index data of the characteristic index at the customer company and index data of the characteristic index at the destination company;
Output means for outputting the customer companies determined to be related by the relationship determination means;
An examination work support device characterized by having the following. The calculation means is
Calculating a plurality of index data based on the transaction data of each of the window dressing destination company and the non-windowing destination company,
The identifying means is
identifying a characteristic index in which the data characteristics of the destination company are expressed from among the plurality of indicators for which index data has been calculated;
The examination work support device according to claim 1, characterized in that: The association determining means is
determining the degree of association between the index data of the characteristic index at the customer company and the index data of the characteristic index at the destination company;
The output means is
Outputting customer companies in descending order of the degree of association;
The examination work support device according to claim 1, characterized in that: The feature index is
Indicator data regarding transaction periodicity or transaction suddenness,
The association determining means is
determining the degree of association based on a variance value of index data of the characteristic index in the customer company and a variance value of index data of the characteristic index in the destination company;
The examination work support device according to claim 3, characterized in that: The feature index is
Indicative data regarding relationships between transaction items,
The association determining means is
determining the degree of association based on a correlation value of index data of the characteristic index at the customer company and a correlation value of index data of the characteristic index at the destination company;
The examination work support device according to claim 3, characterized in that: Examination support equipment at financial institutions
an acquisition procedure for acquiring transaction data of liquidity transaction details of past window dressing companies and non-windowing partner companies;
a calculation procedure for calculating index data based on the transaction data of each of the window dressing company and the non-windowing company;
an identification procedure of identifying a characteristic index in which data characteristics of the window dressing company are expressed based on a data difference between the index data of the window dressing company and the index data of the non-windowing window company;
an acquisition procedure for acquiring transaction data of liquidity transaction details at a customer company;
a calculation procedure for calculating index data of the characteristic index based on the transaction data of the customer company;
a relationship determination procedure for determining a relationship between index data of the characteristic index at the customer company and index data of the characteristic index at the destination company;
an output step of outputting the customer company determined to be related by the relatedness determination step;
A method to support screening operations. to the computer,
an acquisition means for acquiring transaction data of liquidity transaction details of past window dressing companies and non-windowing partner companies;
Calculation means for calculating index data based on the transaction data of each of the window dressing company and the non-windowing company;
Identification means for identifying a characteristic index in which the data characteristics of the window dressing company are expressed based on a data difference between the index data of the window dressing company and the index data of the non-windowing window company;
an acquisition means for acquiring transaction data of liquidity transaction details at a customer company;
Calculation means for calculating index data of the characteristic index based on the transaction data of the customer company;
a relationship determining means for determining a relationship between index data of the characteristic index at the customer company and index data of the characteristic index at the destination company;
Output means for outputting the customer companies determined to be related by the relationship determination means;
An examination support program to make it work.