JP7360808B2 - Banking support system, banking support method and banking support program - Google Patents

Description

The present disclosure relates to a banking support system, a banking support method, and a banking support program for supporting banking operations.

Banks provide various financial services and products such as deposits and loans. Techniques are being considered for predicting the behavior of customers of financial institutions and proposing services and products to them (see, for example, Patent Documents 1 and 2). In the technique described in Patent Document 1, motivation is provided only to customers who are likely to visit the store. The customer behavior management server of Patent Document 1 predicts the date and time of a customer's visit to the store based on the customer's visit history to the store and information on the customer's stop-in locations. It is determined whether the predicted date and time of the customer's visit falls within the busy hours of this store, and when it is determined that the date and time of the customer's visit falls within the busy hours, the customer is notified of information urging him or her to come to the store at another date and time.

Further, the model creation device described in Patent Document 2 models customer behavior based on the customer's awareness of the service as the awareness including awareness of the existence of the service and awareness of the function of the service. The prediction device then uses the customer behavior model to predict which service the customer will select.

Furthermore, techniques for supporting asset management through securities investment by customers and for facilitating lending operations and sound management of financial institutions are also being considered (see, for example, Patent Document 3). In the technology described in Patent Document 3, a financial institution server has a deposit account DB that stores information on customer deposit accounts opened at a financial institution, and a deposit account database that stores information on deposit accounts of customers opened at a financial institution, and a database for issuing securities stored and managed by a securities company. A depository receipt account DB is provided to store information on a customer's depository receipt account opened at a financial institution in such a manner that the financial institution borrows depository receipts from the customer. Then, it receives the customer's buying and selling instructions, creates buying and selling brokerage information, transmits it to the securities company server, receives buying and selling result information from the securities company server, and updates the balances in the deposit account DB and the depositary receipt account DB.

Additionally, techniques for supporting business operations using machine learning and the like are also being considered (see, for example, Patent Document 4). In the technology described in Patent Document 4, a reception unit receives a response to the requested work and a reason for the decision from the contractor. The type classification unit classifies the reason for judgment into types based on the reason for judgment and the answer. The type extraction unit extracts judgment reasons belonging to a predetermined type based on the classified types. The providing unit provides the extracted judgment reason.

Japanese Patent Application Publication No. 2016-173649 Japanese Patent Application Publication No. 2005-352728 Japanese Patent Application Publication No. 2007-47940 Japanese Patent Application Publication No. 2016-177477

The technique described in Patent Document 1 predicts the behavior of each customer, but does not predict customer behavior that contributes to the overall management. Furthermore, the technology described in Patent Document 2 does not take into consideration the control of customer behavior at banks. Although the technology described in Patent Document 3 takes into consideration the sound management of financial institutions, it is not intended to encourage customer behavior.

・The banking support system that solves the above problems is an information storage system that stores financial transaction conditions of financial transactions, customer status information, and history information including financial transaction status with customers for each period. A learning method in which a customer behavior prediction model generated by machine learning is recorded using training data, which is a data set including financial transaction conditions, customer situation information, and financial transaction conditions recorded in the information storage system for the same period. and a control unit connected to the result storage unit . The control unit acquires current financial transaction conditions and customer situation information from the information storage system , and applies the information stored in the learning result storage unit to the current financial transaction conditions and customer situation information. Financial transaction conditions corresponding to the target value of financial transaction status are calculated using a customer behavior prediction model .

According to the present disclosure, accurate and efficient execution of banking operations can be supported.

FIG. 1 is an explanatory diagram of a banking support system according to the present embodiment. An explanatory diagram of a hardware configuration example. FIG. 4 is an explanatory diagram of the information storage unit of the present embodiment, in which (a) is a customer information storage unit, (b) is an account information storage unit, (c) is a transaction information storage unit, and (d) is a condition information storage unit. Explanatory diagram. FIG. 3 is an explanatory diagram of an information storage unit of the present embodiment, in which (a) is an explanatory diagram of an external situation storage unit, (b) is an explanatory diagram of a regulation information storage unit, and (c) is an explanatory diagram of a management information storage unit. An explanatory diagram of the relationship between customer behavior and bank behavior according to the present embodiment. An explanatory diagram of a balance sheet of this embodiment. An explanatory diagram of a customer behavior model according to the present embodiment. FIG. 7 is an explanatory diagram of a processing procedure according to another embodiment. An explanatory diagram of a customer behavior model according to another embodiment. FIG. 7 is an explanatory diagram of a processing procedure according to another embodiment.

(First embodiment)
An embodiment of a banking support system, a banking support method, and a banking support program will be described below with reference to FIGS. 1 to 7. In this embodiment, in order to realize an accurate balance sheet in a bank (financial institution), it is assumed that financial transaction conditions of services and products provided to customers of the bank are controlled.

Here, as shown in FIG. 1, a banking system including a user terminal 10, a support server 20, and a back-end system 40 (business system) is used.
First, the concept of this embodiment will be explained using FIG. 5. Assume a case where customer behavior A1 is predicted using machine learning. Customer behavior A1 includes depositing funds to a bank and borrowing funds from a bank. This customer behavior A1 depends on the customer situation D01 according to the segment to which the customer attribute belongs, and the external situation D02 such as prices and market conditions. In this case, the bank can encourage customer behavior A1 by controlling deposit conditions D03 and lending conditions D04 as financial transaction conditions. Then, the bank's cash deposits D05 are determined by the borrowings (loans) and deposits based on this customer behavior A01.

Bank action A2 is performed in accordance with this cash deposit D05. Banking action A2 includes financing and investment. In this embodiment, it is assumed that the investment is, for example, the purchase of securities. This bank behavior A2 depends on external conditions D11 such as market conditions and interbank interest rates, and regulations D12 such as constraints.

Then, the bank's balance sheet BS1 is determined based on cash and deposits D05, funding, and investments.
Next, the bank's balance sheet BS1 will be explained using FIG. This balance sheet BS1 is composed of cash and deposits, loans, securities, and other items as debits. In addition, the credit side consists of customer deposits and net assets. In this case, each item constituting the balance sheet BS1 has priorities and restrictions.

For example, priority (1) is "cash deposits are a% (for example, about 20%) of customer deposits". Cash and deposits are a provision for the cancellation of customer deposits, but since profitability is low, it is better to have a smaller proportion of cash and deposits.

Priority (2) is ``loans and securities account for b% (for example, about 70%) of customer deposits.'' This is because although loans and securities have high profitability, they involve credit risk, which lowers the equity ratio. Therefore, it is necessary to control the amount by the loan interest rate and loan period.

Priority (3) includes "securities are c% (for example, about 10%) excluding priorities (1) and (2)". Securities are subordinated to loans in terms of profitability, but if they are government bonds, there is no credit risk. Therefore, it will be used as surplus funds in cash and deposits.

Furthermore, as constraint (1), there are regulations regarding the "equity ratio." This capital adequacy ratio regulation is a capital adequacy ratio to risk assets, and is a regulation that sets the Tier 1 ratio of common stocks, etc. to about 10%.

Constraint (2) includes regulations regarding "net assets." The net assets are, for example, about 10% of the total assets. This net asset will be used as a buffer in the event of a loss.
Constraint (3) includes regulations regarding the "liquidity coverage ratio." The liquidity coverage ratio is to secure highly liquid assets in response to capital outflows during 30-day periods of stress.

Constraint (4) is the regulation regarding the "stable procurement ratio." This stable funding ratio ensures stable funding over the medium to long term for assets that are difficult to sell in the short term.
Constraint (5) relates to "IRRBB (bank interest rate risk)". Specifically, the loss from interest rate risk (for example, when interest rates rise by 200 bp for the US dollar and the euro, and 100 bp for the Japanese yen) is to be limited to 15% of Tier 1.

The user terminal 10 is a computer terminal used by a bank representative.
The support server 20 includes a control section 21 and a learning result storage section 22. The control section 21 includes a support management section 211, a learning section 212, a prediction section 213, and an analysis section 214.

The support management unit 211 executes processing to support banking operations. Specifically, the support management section 211 acquires various information from the backend system 40 and provides it to the learning section 212 and the prediction section 213. Furthermore, a prediction result is acquired from the prediction unit 213 and output to the user terminal 10. The support management unit 211 also maintains a condition candidate list. This condition candidate list records data regarding a plurality of deposit condition candidates and loan condition candidates used in support processing.

The learning unit 212 performs machine learning using teacher data consisting of variables constituting the input layer and the output layer to generate a predictive model. This learning unit 212 generates a prediction model used for prediction from various information by multivariate analysis, machine learning, and deep learning.

The prediction unit 213 uses the prediction model generated by the learning unit 212 to execute a process of predicting each variable.
The analysis unit 214 executes processing for predicting actions that will maximize profits in business activities of a company. The analysis unit 214 holds data regarding standard ratios for evaluating cash and deposits. Furthermore, the analysis unit 214 executes investment simulation and procurement simulation according to cash deposits and customer deposit balances. In investment simulations and procurement simulations, simulations are performed taking into account the above-mentioned regulations (priorities and constraints).

In the investment simulation, a simulation is performed in which surplus cash and deposits are used to purchase securities. For example, predict the profit and loss if you purchase government bonds with your current surplus.

In the procurement simulation, a simulation of procuring cash is performed using insufficient cash and deposits. For example, predict the profit and loss when cash is procured using interbank interest rates using current surplus funds.

The learning result storage unit 22 records predictive model data for predicting customer behavior. This predictive model data is recorded when learning is performed based on customer situations, external situations, and interest rates. The learning result storage unit 22 records predictive model data regarding customers.

Backend system 40 functions as an information storage system. Therefore, the back-end system 40 includes a customer information storage section 42, an account information storage section 43, a transaction information storage section 44, a condition information storage section 45, an external situation storage section 46, a regulatory information storage section 47, and a management information storage section 48. Equipped with.

As shown in FIG. 3(a), the customer information storage unit 42 records customer management records regarding bank customers. A customer management record is registered when a customer opens an account at a bank, and is updated when a change is made. The customer management record records data regarding customer identifiers such as customer numbers, customer information, customer attributes, and account identifiers.

The customer identifier data area records data regarding an identifier for identifying each customer of the bank.
In the customer information data area, data regarding the customer's personal information (name, account identifier, contact information, address, etc.) is recorded.

The customer attribute data area records data regarding the customer's attributes (age, gender, occupation, etc.). This customer attribute determines the segment to which the customer belongs. This customer attribute is stored in association with the registration time and change time.

In the account identifier data area, data related to an identifier (head/branch code, type code, account number) for identifying the account held by this customer is recorded.
As shown in FIG. 3(b), the account information storage unit 43 records an account management record 430 regarding a deposit account opened at a bank. The account management record 430 is recorded when the balance status of an account opened at a bank is statistically evaluated. This account management record 430 records data regarding evaluation time, segment, deposit type, and total deposit balance.

The evaluation time data area records data regarding the time (year, month, and day) when an account opened at a bank was evaluated.
The segment data area records data regarding identifiers for specifying segments of customer attributes.

The deposit type data area records data related to an identifier for specifying the type of deposit (for example, ordinary deposit, time deposit, etc.).
The total deposit balance data area records data regarding the total balance of all accounts of deposit types held by customers belonging to this segment.

As shown in FIG. 3C, the transaction information storage unit 44 records a transaction management record 440 regarding a transaction (in this embodiment, a loan) with a customer. Transaction management records are recorded when the status of transactions with customers is statistically evaluated. This transaction management record records data regarding evaluation time, segment, loan type, and total loan balance.

The evaluation time data area records data regarding the time (year, month, and day) when the bank's loan was evaluated.
The segment data area records data regarding identifiers for specifying segments of customer attributes.

In the loan type data area, data related to an identifier for specifying the type of loan (eg, home loan, card loan, etc.) is recorded.
In the total loan balance data area, data regarding the total balance of all loans of the loan type to customers belonging to this segment is recorded.

As shown in FIG. 3(d), the condition information storage section 45 records a deposit condition management record 450 and a loan condition management record 451. The deposit condition management record 450 and the loan condition management record 451 are recorded when financial transaction conditions (deposit conditions and loan conditions) to be provided to a customer are determined.

The deposit condition management record 450 records data regarding change timing, deposit type, and deposit conditions.
In the change time data area, data regarding the year, month, and day when deposit conditions were changed is recorded.

The deposit type data area records data regarding an identifier for specifying the type of deposit.
The deposit condition data area records data regarding changed deposit conditions (deposit interest rate, deposit period).

The lending condition management record 451 records data regarding change timing, lending type, and lending conditions.
In the change time data area, data regarding the date when the loan conditions were changed is recorded.

In the lending type data area, data related to an identifier for specifying the type of lending is recorded. Here, the lending type may be recorded in the lending type data area according to the segment and credit score of the customer to whom the loan is made.

Data regarding changed loan conditions (loan interest rate, loan period) is recorded in the loan condition data area.
As shown in FIG. 4(a), the external situation storage section 46 records data regarding various external situations. In this embodiment, economic index information 461, market information 462, and interest rate information 463 are used as external situation information.

In the economic index information 461, information related to economic indicators is recorded in association with the time (year, month, and day) when the economic index information was acquired. As this economic index, for example, information regarding GDP, prices, etc. can be used.

In the market information 462, market information is recorded in association with the time (year, month, and day) when the market information was acquired. As this market information, for example, index information regarding government bonds (domestic and overseas), stocks (domestic and overseas), foreign exchange, etc. can be used.

In the interest rate information 463, interest rate information is recorded in association with the time (year, month, and day) when the interest rate information was acquired. As this interest rate information, for example, own bank interest rate, overseas interest rate, other bank interest rate, interbank interest rate, etc. can be used.

As shown in FIG. 4(b), the regulation information storage section 47 records regulation information regarding various regulations. This regulation information is recorded when a regulation is registered by a person in charge.
In the restriction information, information regarding the restriction (priority, restrictions, etc.), that is, priority information 471 and restriction information 472, is recorded in association with the time (year, month, and day) when the restriction was set.

As shown in FIG. 4(c), the management information storage unit 48 records various information regarding each element constituting a balance sheet that explains the bank's management status. In this embodiment, held securities information 481, funding information 482, and net worth information 483 are recorded.

In the held securities information 481, information regarding securities held by the bank (brand, quantity, price) is recorded.
In the funding information 482, information regarding funds (source, amount) procured by the bank is recorded.

Net asset information 483 records information regarding net assets (asset details, prices) held by the bank.
(Hardware configuration example)
FIG. 2 shows an example of the hardware configuration of the information processing device H10 that functions as the user terminal 10 to the backend system 40, etc.

The information processing device H10 includes a communication device H11, an input device H12, a display device H13, a storage section H14, and a processor H15. Note that this hardware configuration is just an example, and other hardware may be included.

The communication device H11 is an interface that establishes a communication path with other devices and executes data transmission and reception, and is, for example, a network interface card, a wireless interface, or the like.

The input device H12 is a device that receives input from a user or the like, and is, for example, a mouse, a keyboard, or the like.
The display device H13 is a display, touch panel, or the like that displays various information.

The storage unit H14 is a non-transitory computer-readable medium or storage device (for example, the learning result storage unit 22, advance preparation information storage unit 23, customer information storage unit 42 to management information storage unit 49). Examples of the storage unit H14 include ROM, RAM, hard disk, and the like.

The processor H15 controls each process in the user terminal 10 to the back-end system 40 using programs and data stored in the storage unit H14. Examples of the processor H15 include a CPU, an MPU, and the like. This processor H15 expands a program stored in a ROM or the like into a RAM and executes various processes corresponding to various processing. For example, when the banking support program is activated, the processor H15 operates a process that executes each process shown in FIG. 7, which will be described later.

(Support processing)
The support processing will be explained using FIGS. 7 and 8.
Here, as shown in FIG. 7, a customer behavior prediction model that predicts increases and decreases in borrowings and deposits is calculated from deposit conditions (before and after change), lending conditions (before and after change), external circumstances, and customer conditions.

As shown in FIG. 8, the control unit 21 of the support server 20 executes machine learning processing of customer behavior (step S1-1). Specifically, the support management unit 211 of the control unit 21 specifies, from the condition information storage unit 45, the date on which the deposit interest rate or loan interest rate was changed (interest rate change time) as the evaluation time. Then, the support management unit 211 acquires deposit conditions and lending conditions before and after the interest rate change period from the condition information storage unit 45. Next, the support management unit 211 acquires external situation information at the evaluation time from the external situation storage unit 46. Here, economic index information 461, market information 462, and interest rate information 463 are acquired as external conditions. Furthermore, the support management unit 211 acquires customer attributes (segments) recorded in the customer management record 420 at the evaluation time from the customer information storage unit 42, and calculates the distribution of customer attributes as the customer situation. Next, the support management unit 211 obtains the total account balance amount recorded in the account information storage unit 43 and the total loan balance amount recorded in the transaction information storage unit 44 before and after the evaluation period. Then, the support management unit 211 calculates the increase/decrease (amount) in deposits by subtracting the total account balance before the evaluation time from the total account balance after the evaluation time. Further, the support management unit 211 calculates the increase/decrease in loans (amount) by subtracting the total loan balance before the evaluation period from the total loan balance after the evaluation period.

Then, the support management department 211 creates a data set (teacher data) consisting of deposit conditions (before and after the evaluation period), loan conditions (before and after the evaluation period), external conditions, customer conditions, increases and decreases in borrowings, and increases and decreases in deposits for each evaluation period. do. Next, the learning unit 212 calculates a customer behavior prediction model by performing machine learning using deposit conditions to customer status as an input layer and loan increase/decrease and deposit increase/decrease as an output layer.

Next, the control unit 21 of the support server 20 executes a learning result recording process (step S1-2). Specifically, the learning unit 212 of the control unit 21 records the generated customer behavior prediction model in the learning result storage unit 22.

Next, the control unit 21 of the support server 20 executes a current status setting process (step S1-3). Specifically, the support management section 211 of the control section 21 acquires the current external situation from the external situation storage section 46. Further, the support management unit 211 acquires customer attributes (segments) recorded in the current customer management record 420 from the customer information storage unit 42, and calculates the distribution of customer attributes as the customer situation. The prediction unit 213 then sets the current external situation and customer situation in the input layer of the customer behavior prediction model.

Next, the control unit 21 of the support server 20 executes a temporary setting process for conditions (step S1-4). Specifically, the support management unit 211 of the control unit 21 obtains new deposit condition candidates and loan condition candidates from the condition candidate list for the current deposit conditions and loan conditions. Then, the prediction unit 213 sets the current deposit conditions, lending conditions, and the acquired deposit condition candidates and lending condition candidates in the input layer of the customer behavior prediction model.

Next, the control unit 21 of the support server 20 executes an increase/decrease prediction process (step S1-5). Specifically, the prediction unit 213 of the control unit 21 uses the customer behavior prediction model to calculate the increase/decrease in deposits and increase/decrease in borrowing in the output layer.

Next, the control unit 21 of the support server 20 executes evaluation processing of cash deposits and customer deposits (step S1-6). Specifically, the support management unit 211 of the control unit 21 updates the current balance of customer deposits and cash deposits based on the increase/decrease in deposits and the increase/decrease in borrowing calculated by the prediction unit 213. The analysis unit 214 then calculates the ratio of cash deposits to customer deposits and compares it with the standard ratio.

Here, if the ratio of cash and deposits is equal to or higher than the standard ratio, the control unit 21 of the support server 20 executes an investment simulation process (step S1-7). Specifically, the analysis unit 214 of the control unit 21 obtains information on purchasable securities from an external site, and calculates the funds required for investment. In this case, the investment amount is calculated so as to satisfy various regulations (priorities and constraints) recorded in the regulation information storage unit 47.

On the other hand, if the ratio of cash and deposits is less than the standard ratio, the control unit 21 of the support server 20 executes a procurement simulation process (step S1-8). Specifically, the analysis unit 214 of the control unit 21 acquires available procurement information from an external site and calculates the funds required for procurement. In this case, a procurement simulation is performed to satisfy various regulations (priorities and constraints) recorded in the regulation information storage section 47.

Next, the control unit 21 of the support server 20 executes balance sheet evaluation processing (step S1-9). Specifically, the support management unit 211 of the control unit 21 applies the simulation results to the balance sheet. Further, the support management section 211 sets various information recorded in the management information storage section 48 in the balance sheet. Next, the support management unit 211 applies the simulation results of the investment simulation process (step S1-7) and the procurement simulation process (step S1-8) to the balance sheet. Then, the support management unit 211 temporarily stores the calculated balance sheet in association with deposit conditions and lending conditions.

Next, the control unit 21 of the support server 20 executes a determination process as to whether or not the process is finished (step S1-10). Specifically, the support management unit 211 of the control unit 21 determines that when all deposit condition candidates and loan condition candidates recorded in the condition candidate list are used in the condition provisional setting process (step S1-4), determines that the evaluation process is finished. On the other hand, if there are deposit condition candidates or lending condition candidates that have not yet been used in the condition candidate list, it is determined that the process has not ended.

If it is determined that the process has not ended (“NO” in step S1-10), the control unit 21 of the support server 20 repeats the process starting from the condition provisional setting process (step S1-4).
If it is determined that the process has ended (“YES” in step S1-10), the control unit 21 of the support server 20 executes output processing (step S1-11). Specifically, the support management unit 211 of the control unit 21 identifies a balance sheet that satisfies the priority and constraints among the temporarily stored balance sheets, and outputs the deposit conditions and lending conditions in this case to the user terminal 10. . Then, using the output balance sheet, financial transaction conditions that provide the target value (maximum profit) are specified.

As described above, according to this embodiment, the following effects can be obtained.
(1-1) In the present embodiment, the control unit 21 of the support server 20 executes machine learning processing of customer behavior (step S1-1). For each evaluation period, create a data set (teacher data) consisting of deposit conditions, loan conditions, external conditions, customer conditions, increases and decreases in borrowings, and increases and decreases in deposits. Next, the learning unit 212 calculates a customer behavior prediction model by performing machine learning using deposit conditions to customer status as an input layer and loan increase/decrease, and deposit increase/decrease as an output layer. As a result, by using deposit conditions and lending conditions as an input layer, it is possible to predict increases and decreases in borrowings and deposits when financial transaction conditions are changed.

By using external conditions as an input layer, it is possible to predict increases and decreases in borrowings and deposits in consideration of external conditions.
By using the customer situation as an input layer, it is possible to predict increases and decreases in borrowings and deposits in consideration of the customer situation.

(1-2) In this embodiment, the control unit 21 of the support server 20 executes a temporary setting process of conditions (step S1-4) and a process of predicting increase/decrease (step S1-5). This makes it possible to predict increases and decreases in deposits and loans using financial transaction condition candidates.

(1-3) In the present embodiment, the control unit 21 of the support server 20 executes evaluation processing of cash deposits and customer deposits (step S1-6). If the ratio of cash and deposits is high, the control unit 21 of the support server 20 executes investment simulation processing (step S1-7). On the other hand, if the ratio of cash and deposits is low, the control unit 21 of the support server 20 executes a procurement simulation process (step S1-8). This makes it possible to optimize the balance of cash and deposits.

In this case, simulation is performed so as to satisfy various regulations (priorities and constraints) recorded in the regulation information storage section 47. This makes it possible to perform simulations that take management priorities and banking regulations into consideration.

(1-4) In the present embodiment, the control unit 21 of the support server 20 executes balance sheet evaluation processing (step S1-9) and output processing (step S1-11). This makes it possible to evaluate the impact on the balance sheet of changing financial transaction conditions.

(Second embodiment)
Next, a second embodiment of a banking support system, a banking support method, and a banking support program will be described with reference to FIGS. 9 and 10. Note that the second embodiment has a configuration in which only the customer behavior prediction method of the first embodiment is changed, so detailed explanations of similar parts will be omitted. In the first embodiment, a customer behavior prediction model is calculated by performing machine learning using deposit conditions to customer status as an input layer and loan increase/decrease, and deposit increase/decrease as an output layer.

In the second embodiment, as shown in FIG. 9, deposit conditions (before change), loan conditions (before change), external conditions, and customer conditions are used to obtain the target increase/decrease in borrowings and increase/decrease in deposits. Calculate a customer behavior prediction model that predicts lending conditions (after changes) and lending conditions (after changes).

As shown in FIG. 10, the control unit 21 of the support server 20 executes machine learning processing of customer behavior (step S2-1). Specifically, similar to step S1-1, the support management unit 211 of the control unit 21 determines deposit conditions, lending conditions, external conditions, customer Create a dataset (teacher data) consisting of the situation, increases and decreases in borrowings, and increases and decreases in deposits. Next, the learning unit 212 uses deposit conditions (before change), loan conditions (before change), increase/decrease in borrowings, increase/decrease in deposits, external circumstances, and customer conditions as input layers, and uses deposit conditions (after change), loan conditions (after change) as input layers. ) as the output layer and performs machine learning to calculate a customer behavior prediction model.

First, the control unit 21 of the support server 20 executes a learning result recording process, similar to step S1-2 (step S2-2).
Next, the control unit 21 of the support server 20 executes balance sheet evaluation processing (step S2-3). Specifically, the analysis unit 214 of the control unit 21 acquires various information from the management information storage unit 48 and inputs it into the balance sheet. Next, the analysis unit 214 adjusts each item of the balance sheet within a predetermined fluctuation range to create a target balance sheet (target balance sheet). This target balance sheet is, for example, a balance sheet that maximizes profits while satisfying the priority items and constraints recorded in the regulatory information storage unit 47.

Next, the control unit 21 of the support server 20 executes a target amount setting process for cash and deposits from the balance sheet (step S2-4). Specifically, the analysis unit 214 of the control unit 21 calculates the increase or decrease in deposits from the difference between the customer deposits on the target balance sheet and the customer deposits on the current balance sheet. Furthermore, the analysis unit 214 calculates an increase or decrease in borrowing from the difference between the loan on the target balance sheet and the loan on the current balance sheet.

Next, the control unit 21 of the support server 20 executes the current situation setting process, similar to step S1-3 (step S2-5).
Next, the control unit 21 of the support server 20 executes financial transaction condition prediction processing (step S2-6). Specifically, the prediction unit 213 of the control unit 21 inputs the increase/decrease in deposits and the increase/decrease in borrowing calculated in step S2-4 into the customer behavior prediction model. The prediction unit 213 then calculates deposit conditions and lending conditions using the customer behavior prediction model.

Next, the control unit 21 of the support server 20 executes output processing (step S2-7). Specifically, the support management unit 211 of the control unit 21 outputs the calculated deposit conditions (after change) and lending conditions (after change) to the user terminal 10.

As described above, according to this embodiment, the following effects can be obtained.
(2-1) In the present embodiment, the control unit 21 of the support server 20 executes machine learning processing of customer behavior (step S2-1). Here, the learning unit 212 uses deposit conditions (before change), loan conditions (before change), increase/decrease in borrowings, increase/decrease in deposits, external circumstances, and customer conditions as input layers, and uses deposit conditions (after change), loan conditions (after change) as input layers. ) as the output layer and performs machine learning to calculate a customer behavior prediction model. This makes it possible to predict financial transaction conditions for realizing the target balance sheet.

This embodiment can be modified and implemented as follows. This embodiment and the following modified examples can be implemented in combination with each other within a technically consistent range.
- In the first embodiment, an optimized balance sheet is searched for by changing deposit conditions and lending conditions. Alternatively, machine learning may be used to predict bank behavior in order to calculate a better balance sheet. In this case, bank behavior that associates bank behavior patterns with balance sheet evaluation results is used, using training data that includes bank external situation information, cash and deposits, bank behavior patterns (funding or investment), and balance sheet evaluation results. Generate a predictive model. Furthermore, a bank behavior prediction model may be generated using financial transaction regulation information in the input layer.

- In each of the above embodiments, financial transactions (deposits and borrowings) as customer behavior are predicted using financial transaction conditions. Financial transactions are not limited to deposits and borrowings, and may be combined with either one or other transactions. For example, customer behavior may be predicted using products and services provided by banks. In this case, the backend system 40 is provided with a service information storage unit and a product information storage unit to record historical information regarding the provision conditions and provision status of services and products. Then, a customer behavior model is calculated by performing machine learning using the provision conditions and provision status information of services and products as training data as a data set.

- In each of the above embodiments, teacher data is created using the bank's own information. In addition to this, machine learning may be performed by including information from other banks. Here, information that cannot be obtained from other banks is supplemented using the bank's own information. For example, as for customer information, the location information of the other bank's branch is acquired and substituted with the customer information of the own bank's branch that is located near the other bank's branch.

- In each of the above embodiments, external situation information is recorded in the backend system 40. The external situation information may be obtained from an external system such as an information providing site.
- In each of the above embodiments, the user terminal 10, the support server 20, and the backend system 40 are used. The hardware configuration is not limited to this. For example, these may be configured integrally.

- The control unit 21 is not limited to performing software processing for all processes executed by itself. For example, the control unit 21 may include a dedicated hardware circuit (for example, an application-specific integrated circuit: ASIC) that performs hardware processing for at least part of the processing that the control unit 21 executes. That is, the control unit 21 includes [1] one or more processors that operate according to a computer program (software), [2] one or more dedicated hardware circuits that execute at least some of various processes, or [3] Can be configured as a circuit including a combination thereof. A processor includes a CPU and memory, such as RAM and ROM, where the memory stores program codes or instructions configured to cause the CPU to perform processing. Memory or computer-readable media includes any available media that can be accessed by a general purpose or special purpose computer.

Next, technical ideas that can be understood from the above embodiment and other examples will be additionally described below.
[a] An information storage system that stores financial transaction conditions of financial transactions, customer status information, history information including financial transaction status with customers, and external status information for each period;
A banking business support system comprising a user terminal and a control unit connected to the information storage system,
The control section,
Obtaining external situation information, financial transaction conditions, customer situation information, and financial transaction status for the same period from the information storage system,
Perform machine learning using training data with the acquired information as a dataset to generate a customer behavior prediction model that associates financial transaction conditions and financial transaction status,
A banking business support system, characterized in that the customer behavior prediction model is used to calculate financial transaction conditions corresponding to a target value of the transaction situation with respect to the current external situation and the customer situation.
[b] The banking business support system according to [a], wherein an interest rate is used as the financial transaction condition.
[c] The financial transaction includes at least one of deposits and loans;
The banking business support system according to [a] or [b], characterized in that the business situation of the bank is predicted based on the deposit situation and lending situation as the financial transaction situation.
[d] further comprising a management information storage unit for storing the bank's asset status;
The control section,
Obtaining the amount of financing and investment calculated based on the asset status stored in the management information storage unit,
The banking business support system according to any one of [a] to [c], characterized in that the balance sheet is calculated based on the amounts of the financing and investment.
[e] The control section,
Obtaining a target balance sheet, identifying the amount of financing or investment in the target balance sheet,
The banking support system according to any one of [a] to [d], characterized in that the target value of the financial transaction status is calculated based on the amount.
[f] The control unit generates a bank behavior prediction model that associates bank behavior patterns and balance sheets using training data including bank external situation information, cash deposits, bank behavior patterns, and balance sheet evaluation results. The banking business support system according to claim [e], characterized in that:
[g] The information storage system further stores regulatory information on financial transactions;
The banking business support according to any one of [a] to [f], wherein the control unit predicts banking behavior using financial transaction regulation information stored in the information storage system. system.
[h] An information storage system that stores financial transaction conditions of financial transactions, customer status information, history information including financial transaction status with customers, and external status information for each period;
A method for supporting banking operations using a banking operations support system comprising a user terminal and a control unit connected to the information storage system, the method comprising:
The control section,
Obtaining external situation information, financial transaction conditions, customer situation information, and financial transaction status for the same period from the information storage system,
Perform machine learning using training data with the acquired information as a dataset to generate a customer behavior prediction model that associates financial transaction conditions and financial transaction status,
A banking business support method comprising calculating financial transaction conditions corresponding to a target value of a transaction situation using the customer behavior prediction model with respect to current external conditions and customer conditions.
[i] An information storage system that stores financial transaction conditions of financial transactions, customer status information, history information including financial transaction status with customers, and external status information for each period;
A program for supporting banking operations using a banking operations support system comprising a user terminal and a control unit connected to the information storage system, the program comprising:
The control section,
Obtaining external situation information, financial transaction conditions, customer situation information, and financial transaction status for the same period from the information storage system,
Perform machine learning using training data with the acquired information as a dataset to generate a customer behavior prediction model that associates financial transaction conditions and financial transaction status,
A banking business support program characterized in that the program functions as a means for calculating financial transaction conditions corresponding to a target value of a transaction situation using the customer behavior prediction model with respect to current external conditions and customer conditions.

DESCRIPTION OF SYMBOLS 10... User terminal, 20... Support server, 21... Control unit, 22... Learning result storage unit, 211... Support management unit, 212... Learning unit, 213... Prediction unit, 214... Analysis unit, 40... Back end system, 42 ...Customer information storage section, 43... Account information storage section, 44... Transaction information storage section, 45... Condition information storage section, 46... External situation storage section, 47... Regulatory information storage section, 48... Management information storage section.

Claims (7)

an information storage system that stores financial transaction conditions of financial transactions, customer status information, history information including financial transaction status with customers , and external status information for each period;
A customer behavior prediction model generated by machine learning is recorded using training data as a data set including financial transaction conditions, customer situation information, financial transaction situation, and external situation information recorded in the information storage system during the same period. A banking business support system comprising a learning result storage unit and a control unit connected to the
The control section,
Obtaining current financial transaction conditions, customer situation information, and external situation information from the information storage system;
Calculate financial transaction conditions corresponding to a target value of the financial transaction situation using the customer behavior prediction model recorded in the learning result storage unit for the current financial transaction conditions, customer situation information, and external situation information. A banking support system that is characterized by: 2. The banking business support system according to claim 1, wherein an interest rate is used as the financial transaction condition. It is further equipped with a management information storage unit that stores the bank's asset status.
The control section,
Obtaining the amount of financing and investment calculated based on the asset status stored in the management information storage unit,
3. The banking business support system according to claim 1, wherein a balance sheet is calculated based on the amount of funding and investment. The control section,
A target balance sheet and a current balance sheet including customer deposits and loans are obtained, the target balance sheet and the current balance sheet are compared, and the amount of financing is calculated from the difference in the customer deposits. At the same time, calculate the investment amount from the difference between the loans,
The banking business support according to any one of claims 1 to 3 , wherein the financial transaction conditions are calculated using the customer behavior prediction model for the difference in the customer deposit and the difference in the loan. system. The control unit generates a bank behavior prediction model that associates bank behavior patterns with balance sheet evaluation results using teacher data including bank external situation information, cash deposits, bank behavior patterns, and balance sheet evaluation results. The banking business support system according to claim 4 , characterized in that: an information storage system that stores financial transaction conditions of financial transactions, customer status information, history information including financial transaction status with customers , and external status information for each period;
A customer behavior prediction model generated by machine learning is recorded using training data as a data set including financial transaction conditions, customer situation information, financial transaction situation, and external situation information recorded in the information storage system during the same period. A method for supporting banking operations using a banking operations support system comprising a learning result storage section and a control section connected to the
The control section,
Obtaining current financial transaction conditions, customer situation information, and external situation information from the information storage system;
Calculate financial transaction conditions corresponding to a target value of the financial transaction situation using the customer behavior prediction model recorded in the learning result storage unit for the current financial transaction conditions, customer situation information, and external situation information. A banking business support method characterized by: an information storage system that stores financial transaction conditions of financial transactions, customer status information, history information including financial transaction status with customers , and external status information for each period;
A customer behavior prediction model generated by machine learning is recorded using training data as a data set including financial transaction conditions, customer situation information, financial transaction situation, and external situation information recorded in the information storage system during the same period. A program for supporting banking operations using a banking operations support system comprising a learning result storage section and a control section connected to the
The control section,
Obtaining current financial transaction conditions, customer situation information, and external situation information from the information storage system;
Calculate financial transaction conditions corresponding to a target value of the financial transaction situation using the customer behavior prediction model recorded in the learning result storage unit for the current financial transaction conditions, customer situation information, and external situation information. A banking support program that is characterized by functioning as a means to

Images