CN111932018B

CN111932018B - Bank business performance contribution information prediction method and device

Info

Publication number: CN111932018B
Application number: CN202010812638.0A
Authority: CN
Inventors: 郑洁锋; 温丽明; 帅翡芍
Original assignee: Industrial and Commercial Bank of China Ltd ICBC
Current assignee: Industrial and Commercial Bank of China Ltd ICBC
Priority date: 2020-08-13
Filing date: 2020-08-13
Publication date: 2023-09-19
Anticipated expiration: 2040-08-13
Also published as: CN111932018A

Abstract

The invention provides a banking business performance contribution information prediction method and device, which can be used in the financial field or other fields. The method comprises the following steps: acquiring service data; acquiring asset business data, liability business data and corresponding characteristic information through data processing; training a first machine learning model which is built in advance by taking the obtained data as a training sample to obtain an initial asset liability configuration model and a corresponding training result; training the second machine learning model by taking the training result as a training sample to obtain an initial internal funds transfer cost pricing model; and taking the training result of the initial internal funds transfer cost pricing model as a calculation parameter, and calculating by using a rolling algorithm to obtain service performance contribution information. According to the invention, the characteristic value extraction, the machine learning training and the model calculation are carried out on the banking data, so that the prediction of the banking performance contribution information and the optimization of the overall structure of the banking asset liability are realized.

Description

Bank business performance contribution information prediction method and device

Technical Field

The invention relates to the technical field of banking business data processing, in particular to a banking business performance contribution information prediction method and device.

Background

The overall configuration of the commercial banking liabilities, internal funds transfer cost pricing and business performance contribution prediction are several important indicators of commercial banking analysis. The overall configuration condition of the commercial bank asset liabilities, including the corresponding condition of the commercial bank asset of each period and each period liabilities, is an important index and tool for analyzing the structure of the commercial bank asset liabilities. Good liability structures are also a requirement for regulatory compliance. The cost pricing of money transfer in the commercial bank is the cost pricing in the commercial bank and is an important index and tool for measuring and calculating the performance contribution of business of the commercial bank. The business performance contribution of the commercial bank is the difference between the business income and the cost pricing according to the fund transfer cost pricing, and is an important index and tool for counting the customer contribution, analyzing the customer group, internally checking and restoring and the like.

In the prior art, a manual mode is generally adopted by banks for measuring and calculating the operation indexes, various business reports of the liability are collected, classified, summarized, analyzed and measured, the timeliness is low, the accuracy is in doubt, and the measurement and calculation results of the indexes cannot be effectively utilized to guide the bank operation analysis.

Disclosure of Invention

The embodiment of the invention mainly aims to provide a banking business performance contribution information prediction method and device, which realize automatic prediction of business performance contribution information according to the overall configuration condition of commercial banking asset liabilities.

In order to achieve the above object, an embodiment of the present invention provides a banking performance contribution information prediction method, including:

acquiring service data of a bank application service system, wherein the service data comprises expired service data and existing service data;

performing data processing on the service data to obtain asset service data, liability service data and characteristic information corresponding to the asset service data and the liability service data;

training a first machine learning model established in advance by taking the expiration service data, the existing service data and the characteristic information as training samples to obtain an initial asset liability configuration model and a corresponding training result;

training a second machine learning model which is built in advance by taking a training result of the initial asset liability configuration model as a training sample to obtain an initial internal funds transfer cost pricing model and a corresponding training result;

And taking the training result of the initial internal funds transfer cost pricing model as a calculation parameter, and calculating by using a rolling algorithm to obtain service performance contribution information.

Optionally, in an embodiment of the present invention, the data processing on the service data includes: and carrying out supplementary recording on the missing service data and processing special rule data in the service data.

Optionally, in an embodiment of the present invention, the performing data processing on the service data to obtain asset service data and liability service data includes: and classifying the expired service data and the existing service data to obtain asset service data and liability service data.

Optionally, in an embodiment of the present invention, the data processing for the service data to obtain feature information corresponding to the asset service data and liability service data further includes: and screening the expired service data and the existing service data, and extracting characteristic information, wherein the characteristic information comprises customer basic information, service product information, currency information, principal information, damage and benefit information, re-pricing deadline information and residual deadline information.

Optionally, in an embodiment of the present invention, training the pre-established first machine learning model by using the expiration service data, the existing service data and the feature information as training samples, to obtain an initial asset liability configuration model and a corresponding training result includes: and training a pre-established first machine learning model by taking the expiration service data, the existing service data and the re-pricing deadline information as training samples to obtain an initial asset liability configuration model in a re-pricing deadline mode and a corresponding training result.

Optionally, in an embodiment of the present invention, training the pre-established first machine learning model by using the expiration service data, the existing service data and the feature information as training samples, to obtain an initial asset liability configuration model and a corresponding training result further includes: and training a first machine learning model which is built in advance by taking the expiration service data, the existing service data and the residual deadline information as training samples to obtain an initial asset liability configuration model under the residual deadline mode and a corresponding training result.

Optionally, in an embodiment of the present invention, after training the pre-established first machine learning model by using the expiration service data, the existing service data and the feature information as training samples, the training method further includes: and optimizing the asset liability configuration model in the initial re-pricing deadline mode and the asset liability configuration model in the initial residual deadline mode by taking the acquired newly-added business data as a test sample, and generating an asset liability configuration model in the trained re-pricing deadline mode and a corresponding training result, and an asset liability configuration model in the trained residual deadline mode and a corresponding training result.

Optionally, in an embodiment of the present invention, the training result of the initial internal funds-transfer cost pricing model includes internal funds-transfer cost pricing information in a re-pricing deadline mode, where the internal funds-transfer cost pricing information is generated by: and training the initial internal funds transfer cost pricing model by taking the training result of the asset liability configuration model in the initial re-pricing deadline mode as a training sample to obtain the internal funds transfer cost pricing information in the re-pricing deadline mode.

Optionally, in an embodiment of the present invention, the training result of the initial internal funds-transfer cost pricing model includes internal funds-transfer cost pricing information in a remaining period mode, and the internal funds-transfer cost pricing information in the remaining period mode is generated by: and training the initial internal funds transfer cost pricing model by taking the training result of the asset liability configuration model in the initial remaining period mode as a training sample to obtain the internal funds transfer cost pricing information in the remaining period mode.

Optionally, in an embodiment of the present invention, the calculating, using the rolling algorithm to obtain the business performance contribution information, using a training result of the initial internal funds transfer cost pricing model as a calculation parameter includes: and obtaining service performance contribution information in the re-pricing deadline mode through rolling calculation according to the internal funds transfer cost pricing information in the re-pricing deadline mode.

Optionally, in an embodiment of the present invention, the calculating, using the rolling algorithm to obtain the business performance contribution information, using a training result of the initial internal funds transfer cost pricing model as a calculation parameter includes: and obtaining service performance contribution information in the remaining period mode through rolling calculation according to the internal funds transfer cost pricing information in the remaining period mode.

The embodiment of the invention also provides a banking business performance contribution information prediction device, which comprises:

the system comprises a data acquisition unit, a bank application service system and a bank application service system, wherein the data acquisition unit is used for acquiring service data of the bank application service system, and the service data comprises expired service data and existing service data;

the data processing unit is used for carrying out data processing on the business data to obtain asset business data, liability business data and characteristic information corresponding to the asset business data and the liability business data;

the configuration model unit is used for training a first machine learning model which is built in advance by taking the expiration service data, the existing service data and the characteristic information as training samples to obtain an initial asset liability configuration model and a corresponding training result;

the pricing model unit is used for training a second machine learning model which is built in advance by taking a training result of the initial asset liability configuration model as a training sample to obtain an initial internal fund transfer cost pricing model and a corresponding training result;

And the contribution prediction unit is used for calculating and obtaining service performance contribution information by using the rolling algorithm by taking the training result of the initial internal funds transfer cost pricing model as a calculation parameter.

Optionally, in an embodiment of the present invention, the data processing unit is specifically configured to perform complement on missing service data and process special rule data in the service data.

Optionally, in an embodiment of the present invention, the data processing unit is further specifically configured to classify the expiration service data and the existing service data to obtain asset service data and liability service data.

Optionally, in an embodiment of the present invention, the data processing unit is further specifically configured to filter the expiration service data and the existing service data, and extract feature information, where the feature information includes customer basic information, service product information, currency information, principal information, profit-loss information, re-pricing deadline information and remaining deadline information.

Optionally, in an embodiment of the present invention, the configuration model unit is specifically configured to train a first machine learning model built in advance by using the expiration service data, the existing service data and the rendition deadline information as training samples, to obtain an asset liability configuration model in an initial rendition deadline mode and a corresponding training result.

Optionally, in an embodiment of the present invention, the configuration model unit is further specifically configured to train a first machine learning model built in advance by using the expiration service data, the existing service data and the remaining term information as training samples, to obtain an initial balance configuration model in a remaining term mode and a corresponding training result.

Optionally, in an embodiment of the present invention, the configuration model unit is further specifically configured to optimize the asset liability configuration model in the initial re-pricing deadline mode and the asset liability configuration model in the initial remaining deadline mode by using the acquired newly-added service data as a test sample, and generate the asset liability configuration model in the trained re-pricing deadline mode and a corresponding training result, and the asset liability configuration model in the trained remaining deadline mode and a corresponding training result.

Optionally, in an embodiment of the present invention, the pricing model unit is specifically configured to use a training result of the asset liability configuration model in the initial re-pricing deadline mode as a training sample, and obtain the internal funds transfer cost pricing information in the re-pricing deadline mode through training of the initial internal funds transfer cost pricing model.

Optionally, in an embodiment of the present invention, the pricing model unit is further specifically configured to use a training result of the asset liability configuration model in the initial remaining period mode as a training sample, and obtain the internal funds transfer cost pricing information in the remaining period mode through training of the initial internal funds transfer cost pricing model.

Optionally, in an embodiment of the present invention, the contribution prediction unit is specifically configured to obtain, through a rolling calculation, service performance contribution information in the re-pricing deadline mode according to the internal funds transfer cost pricing information in the re-pricing deadline mode.

Optionally, in an embodiment of the present invention, the contribution prediction unit is further specifically configured to obtain, according to the internal funds transfer cost pricing information in the remaining deadline mode, service performance contribution information in the remaining deadline mode through a rolling calculation.

The invention also provides an electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, the processor implementing the above method when executing the program.

The present invention also provides a computer readable storage medium storing a computer program for executing the above method.

According to the invention, through extracting the characteristic value of the banking data, training by machine learning and calculating the model, the prediction of the banking performance contribution information is realized, the reasonable configuration of assets and liabilities of the banking product is supported, the optimization of the overall structure of the banking asset liabilities is realized, the assessment and restoration in the bank are facilitated, and the customer contribution is estimated.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a flowchart of a banking performance contribution information prediction method according to an embodiment of the present invention;

FIG. 2 is a schematic diagram of a decision tree in an embodiment of the present invention;

fig. 3 is a schematic structural diagram of a banking performance contribution information prediction apparatus according to an embodiment of the present invention;

fig. 4 is a schematic structural diagram of a banking performance contribution information prediction system according to an embodiment of the present invention;

Fig. 5 is a schematic structural diagram of a bank data acquisition module according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of a training module for asset liability configuration in an embodiment of the present invention;

FIG. 7 is a schematic diagram of an internal funds-transfer cost-pricing training module according to an embodiment of the invention;

fig. 8 is a schematic structural diagram of a service performance contribution information prediction module in an embodiment of the present invention;

FIG. 9 is a schematic diagram of a data display module according to an embodiment of the present invention;

fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the invention.

Detailed Description

The embodiment of the invention provides a banking business performance contribution information prediction method and device. It should be noted that, the banking performance contribution prediction method and apparatus of the present invention may be used in the financial field, and may also be used in any field other than the financial field.

The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.

Fig. 1 is a flowchart of a banking performance contribution information prediction method according to an embodiment of the present invention, where the method includes:

step S1, service data of a bank application service system is obtained, wherein the service data comprises expired service data and existing service data.

Wherein acquiring the expiration service data and the existing service data of the banking application service system comprises: acquiring service data which are expired in the current year and service data which are not expired currently by a bank application service system; and screening, classifying and processing the service data to obtain asset service data and liability service data and characteristic information corresponding to the asset service data and the liability service data.

And step S2, carrying out data processing on the service data to obtain asset service data, liability service data and characteristic information corresponding to the asset service data and the liability service data.

The service data is acquired through each application server of the bank, and the service data of the application server of the bank comprises but is not limited to: deposit system business data, loan system business data, trade financing system business data, financial market system business data, and clearing system business data.

Specifically, the business data is classified to obtain asset business data and liability business data. And screening the service data, and extracting the characteristic information such as the basic information of clients, the service product information, the currency information, principal information, loss information, the deadline information (the date of information, the expiration date, the date of re-pricing) and the like. Special data processing is carried out on the service data, which comprises the following steps: the missing data is subjected to supplementary recording, and on one hand, key characteristic information missing in the system is supplemented; on the other hand, the information which is not supported by the system to be recorded is subjected to supplementary recording, for example, the rights and interests of owners remained after the liabilities are matched with the assets; and processing special rule data, such as special scenes of loan selling, negative interest rate and the like, and judging and processing.

And step S3, training a pre-established first machine learning model by taking the initial expiration service data, the existing service data and the characteristic information as training samples to obtain an initial asset liability configuration model and a corresponding training result.

And training a pre-established asset liability configuration machine learning model by taking the re-pricing deadline information in the deadline information of the service data, the existing service data and the characteristic information as training samples to obtain an initial re-pricing deadline mode asset liability configuration model. And training a pre-established asset liability configuration machine learning model by taking the remaining deadline information in the deadline information of the current-year expiration service data, the existing service data and the characteristic information as training samples to obtain an initial remaining deadline model asset liability configuration training model.

Further, the newly added business data is used as a new test sample to optimize the machine learning model after initial training, and the trained machine learning model is generated, so that the autonomous adjustment and optimization of the machine learning model are realized. The service data which expires in the current year and the service data which does not expire (existing service data) are taken as initial samples, an initial asset liability configuration model is trained and obtained, the calculation times and the data quantity of the test samples are adjusted through continuous input of the newly added service data, and the deviation is continuously optimized to obtain each leaf node of the decision tree, as shown in figure 2.

The training result of the asset liability configuration model comprises the following steps: the business currencies under the re-pricing deadline mode, the asset deadline codes correspond to the configuration amount of the liability deadline codes and the owner rights and interests information; and under the residual period mode, each business currency, each asset period code corresponds to the configuration amount of each liability period code and the owner rights and interests information.

And S4, training a second machine learning model which is built in advance by taking the training result of the initial asset liability configuration model as a training sample to obtain an initial internal fund transfer cost pricing model and a corresponding training result.

The training result of the internal funds transfer cost pricing model comprises internal funds transfer cost pricing information of each deadline code corresponding to each business currency in a re-pricing deadline mode; and the fund transfer cost pricing information in each time limit code corresponding to each business currency in the remaining time limit mode.

Specifically, according to the training result of the asset liability configuration model as a training sample, the internal funds transfer cost pricing information under the re-pricing deadline mode is obtained through model training. And training according to the training result of the asset liability configuration model as a training sample, and obtaining the internal funds transfer cost pricing information under the residual period mode through model training. And synchronously adjusting the calculation times and the data quantity of the test sample according to the training result of the continuously optimized and adjusted asset liability configuration model as a training sample, and optimizing the deviation to obtain a regression equation, thereby obtaining the internal funds transfer cost pricing model.

Step S3 and step S4 build relatively independent machine learning models in advance according to different machine learning algorithms. And step S3, training the machine learning model according to the expired service data and the existing service data as samples, and continuously optimizing and adjusting the machine learning model according to the newly added service data as samples. And step S4, training and continuously optimizing and adjusting the machine learning model according to the training result of the step S3 as a training sample. For a user, the two trained machine training learning models can input a set of newly-added business data to obtain two training models and obtain the effect of two important bank operation indexes.

On the other hand, each operation index of the bank should be a reaction of the whole business data, and each operation index cannot be contradicted. Therefore, the input of the model training in the step S4 adopts the result value obtained by the model training in the step S3 to be reasonable and reliable.

After training and adding new business data as training samples, the two machine training models after continuous optimization and adjustment can keep synchronous and continuously approach to actual conditions.

And S5, taking the training result of the internal funds transfer cost pricing model as a calculation parameter, and calculating by using a rolling algorithm to obtain service performance contribution information.

Specifically, according to the internal funds transfer cost pricing information in the re-pricing deadline mode, business performance contribution information in the re-pricing deadline mode is obtained through rolling calculation. And obtaining service performance contribution information in the residual deadline mode through rolling calculation according to the internal funds transfer cost pricing information in the residual deadline mode.

As an embodiment of the present invention, the data processing of the service data includes: and carrying out supplementary recording on the missing service data and processing special rule data in the service data.

The method comprises the steps of carrying out supplementary recording on missing data, supplementing key characteristic information missing in a service system, and carrying out supplementary recording on information which is not supported to be recorded by the service system. For example, liabilities match the rights of owners remaining after the asset is completed, and there is often no information about the rights of owners in the business data details, and then an open interface is required to provide a key information complement function. And processing special rule data, such as special scenes of loan selling, negative interest rate and the like, and judging and processing.

As an embodiment of the present invention, performing data processing on the service data to obtain asset service data and liability service data includes: and classifying the expired service data and the existing service data to obtain asset service data and liability service data.

The method comprises the steps of judging whether the service data are due or not according to the service data, wherein the due service data and the existing service data are further distinguished according to different classifications of the due service data and the existing service data, and classifying the due service data and the existing service data into asset service data and liability service data.

As an embodiment of the present invention, performing data processing on the service data to obtain feature information corresponding to the asset service data and liability service data further includes: and screening the expired service data and the existing service data, and extracting characteristic information, wherein the characteristic information comprises customer basic information, service product information, currency information, principal information, damage and benefit information, re-pricing deadline information and residual deadline information.

The characteristic value (characteristic information) includes, among others, asset liability information, customer information (customer number, customer name), business product information (business product type, asset liability identification), currency information (business currency, interest currency), principal information (business principal), profit-and-loss information (business interest rate, business interest), term information (date of interest, date of due, date of re-price, remaining term, re-price term), and the like. As shown in table 1, the attribute values of the present embodiment and the corresponding attribute value table are shown.

TABLE 1

Preprocessing the characteristic value, specifically including: assets and liability deadlines are divided and categorized as shown in tables 2 and 3.

TABLE 2

TABLE 3 Table 3

Liability deadline code	Liability deadline range
		1M	>=1 and<30
3M	>=30 and<90
		6M	>=90 and<180
1Y	>=180 and<360
		3Y	>=360 and<1080
5Y	>=1080 and<1800
		5Y+	>＝1800

and respectively summarizing the business principal and the business interest according to the business currency, the asset liability identification, the classified asset deadline codes and the liability deadline codes.

In this embodiment, training the pre-established first machine learning model by using the expiration service data, the existing service data and the feature information as training samples, to obtain an initial liability configuration model and a corresponding training result includes: and training a pre-established first machine learning model by taking the expiration service data, the existing service data and the re-pricing deadline information as training samples to obtain an initial asset liability configuration model in a re-pricing deadline mode and a corresponding training result.

In this embodiment, training the pre-established first machine learning model by using the expiration service data, the existing service data and the feature information as training samples, to obtain an initial liability configuration model and a corresponding training result further includes: and training a first machine learning model which is built in advance by taking the expiration service data, the existing service data and the residual deadline information as training samples to obtain an initial asset liability configuration model under the residual deadline mode and a corresponding training result.

In this embodiment, training the first machine learning model established in advance by using the expiration service data, the existing service data and the feature information as training samples, to obtain an initial liability configuration model and a corresponding training result, further includes: and optimizing the asset liability configuration model in the initial re-pricing deadline mode and the asset liability configuration model in the initial residual deadline mode by taking the acquired newly-added business data as a test sample, and generating an asset liability configuration model in the trained re-pricing deadline mode and a corresponding training result, and an asset liability configuration model in the trained residual deadline mode and a corresponding training result.

Further, the newly added business data is used as a new test sample to optimize the machine learning model after initial training, and the trained machine learning model is generated, so that the autonomous adjustment and optimization of the machine learning model are realized. The service data which expires in the current year and the service data which does not expire (existing service data) are taken as initial samples, an initial asset liability configuration model is trained and obtained, the calculation times and the data quantity of the test samples are adjusted through continuous input of the newly added service data, and the deviation is continuously optimized to obtain each leaf node of the decision tree, so that the asset liability configuration model is continuously adjusted and optimized.

Specifically, the result information of each leaf node is output according to the characteristic value data and the judging conditions generated by processing through a machine learning decision tree algorithm.

The characteristic value data generated by the input processing includes: business currency, asset liability identification, categorized asset deadline codes and liability deadline codes, summarized business principal and business interest.

As shown in fig. 2, the decision tree decision conditions and decision results include:

judging: deadline type

Determination result 1: period of re-pricing

Determination result 2: residual period of time

Judgment 1: period N liability > period 1 asset-period 1-N-1 liability VS period 1 asset configuration amount

Determination result Yes: period 1 asset-period 1-N liability VS period 1 asset configuration amount

Determination result No: deadline N liability = configured amount

Judging 2: period 1 asset-period 1 to N liability VS period 1 asset configuration amount > asset period 2 amount-period 1 to N-1 liability VS period 2 asset configuration amount

Determination result Yes: asset period 2 amount-period 1-N liability VS period 2 asset configuration amount

Determination result No: deadline 1 asset-deadline 1-N liability VS deadline 1 asset configuration amount = configuration amount

Judging 3: asset period 2 amount-period 1-N liability VS period 2 asset configuration amount > asset period 3 amount-period 1-N-1 liability VS period 3 asset configuration amount

Determination result Yes: asset period 3 amount-period 1-N liability VS period 3 asset configuration amount

Determination result No: asset period 2 amount-period 1-N liability VS period 2 asset configuration amount = configuration amount

Judging 4: asset period Mamount-period 1-N liability VS period M asset configuration amount >0 and asset has no amount greater than period M

Determination result Yes: asset period M amount-period 1-N liability VS period M asset configuration amount = owner equity

Determination result No: assignment 0

The business currencies under the re-pricing deadline mode, the asset deadline codes correspond to the configuration amount of the liability deadline codes and the owner rights and interests information; and under the residual period mode, each business currency, each asset period code corresponds to the configuration amount of each liability period code and the owner rights and interests information. Wherein, the configuration amount is explained in detail as: and if the liability amount of the term N is greater than or equal to the remaining amount of the asset, assigning the remaining amount of the asset, and if the liability amount is less than the remaining amount of the asset, assigning the liability amount. Period M asset remaining amount = period M asset amount-all configured amounts between liability amounts less than liability period N and period M asset amount. When the term N is 1, the term M asset remaining amount=the term M asset amount. The configured amount may specify a period of the liability amount and a period of the asset amount at the same time. For example, if the amount of liabilities with a period of 1M is 100 tens of thousands, the amount of liabilities with a period of 3M is 100 tens of thousands, the amount of assets with a period of 1M is 150 tens of thousands, and the amount of assets with a period of 3M is 150 tens of thousands, the amount of liabilities with a period of 1M and the amount of assets with a period of 1M are 100 tens of thousands; the amount of configuration of the period 3M liability and the period 1M asset is 50 ten thousand; the amount of configuration of the period 3M liabilities and period 3M assets is 50 tens of thousands.

Wherein, the term N, the term M, the term 1, the term 2 and the term 3 represent certain term code information.

In this embodiment, the training result of the initial internal funds-transfer cost pricing model includes internal funds-transfer cost pricing information in a re-pricing deadline mode, which is generated by: and training the initial internal funds transfer cost pricing model by taking the training result of the asset liability configuration model in the initial re-pricing deadline mode as a training sample to obtain the internal funds transfer cost pricing information in the re-pricing deadline mode.

In this embodiment, the training results of the initial internal funds-transfer cost pricing model include internal funds-transfer cost pricing information in a remaining deadline mode, which is generated by: and training the initial asset liability configuration model under the residual deadline mode through the initial internal funds transfer cost pricing model to obtain the internal funds transfer cost pricing information under the residual deadline mode.

Specifically, the result of the training of the asset liability configuration model is transmitted to the internal funds transfer cost pricing model and is used as a training sample and characteristic value data for input.

By multiple linear regression supporting the vector machine regression (SVR) algorithm, by giving a new input sample x, it is deduced from the given data sample what its corresponding output Y is, which is a real number. Regression problems can be described in mathematical language as:

a given set of data samples is { (x) _i ,y _i )|x _i ∈R ⁿ ,y _i E R, i=1, 2,3. Find R ⁿ The last function f (x), a regression equation is derived to infer the y value for any x input with y=f (x).

The y value of the machine learning training result is that the internal fund transfer cost pricing of each deadline code corresponding to each business currency in the re-pricing deadline mode; and pricing the internal funds transfer cost of each deadline code corresponding to each business currency in the residual deadline mode.

The specific implementation is as follows: period M internal funds transfer cost pricing y = period M asset VS period 1 liability configuration amount x period 1 interest rate + period M asset VS period 2 liability configuration amount x period 2 interest rate + period M asset VS period 3 liability configuration amount x period 3 interest rate.

The finer the asset deadline code and liability deadline code presets, the smoother the outgoing internal funds transfer cost curve.

In this embodiment, calculating the service performance contribution information using the rolling algorithm using the training result of the initial internal funds transfer cost pricing model as the calculation parameter includes: and obtaining service performance contribution information in the re-pricing deadline mode through rolling calculation according to the internal funds transfer cost pricing information in the re-pricing deadline mode.

In this embodiment, calculating the service performance contribution information using the rolling algorithm using the training result of the initial internal funds transfer cost pricing model as the calculation parameter includes: and obtaining service performance contribution information in the remaining period mode through rolling calculation according to the internal funds transfer cost pricing information in the remaining period mode.

Specifically, the result of the internal funds transfer cost pricing model training is used as the business performance contribution prediction calculation input parameter. And obtaining performance contribution values of each business by utilizing a rolling algorithm according to the internal funds transfer cost pricing of each term obtained through training of the internal funds transfer cost pricing model.

The specific rolling algorithm is as follows:

service performance contribution value for the re-pricing deadline = service annual rate of return-internal funds transfer cost pricing for the corresponding re-pricing deadline;

remaining term business performance contribution value = business annual rate of return-internal funds transfer cost pricing for the corresponding remaining term.

Further, the service performance contribution prediction result is displayed.

According to the invention, through extracting the characteristic value, training by machine learning and calculating the model of the banking data, the prediction of the banking performance contribution is realized, the reasonable configuration of assets and liabilities of the banking product is supported, the optimization of the overall structure of the banking asset liabilities is realized, the assessment and restoration in the bank are facilitated, and the customer contribution is estimated.

Fig. 3 is a schematic structural diagram of a banking performance contribution information prediction apparatus according to an embodiment of the present invention, where the apparatus includes:

a data acquisition unit 10 for acquiring service data of a banking application service system, the service data including expired service data and existing service data;

the data processing unit 20 is configured to perform data processing on the service data to obtain asset service data, liability service data, and feature information corresponding to the asset service data and the liability service data;

a configuration model unit 30, configured to train a first machine learning model built in advance by using the expiration service data, the existing service data and the feature information as training samples, so as to obtain an initial liability configuration model and a corresponding training result;

a pricing model unit 40, configured to train a second machine learning model that is built in advance according to a training result of the initial asset liability configuration model as a training sample, so as to obtain an initial internal funds transfer cost pricing model and a corresponding training result;

and the contribution prediction unit 50 is configured to calculate, using a rolling algorithm, the business performance contribution information by using the training result of the initial internal funds transfer cost pricing model as a calculation parameter.

As an embodiment of the present invention, the data processing unit is specifically configured to perform the complement on the missing service data and process the special rule data in the service data.

As an embodiment of the present invention, the data processing unit is further specifically configured to classify the expiration service data and the existing service data to obtain asset service data and liability service data.

As an embodiment of the present invention, the data processing unit is further specifically configured to filter the expiration service data and the existing service data, and extract feature information, where the feature information includes customer basic information, service product information, currency information, principal information, damage information, re-pricing deadline information and remaining deadline information.

In this embodiment, the configuration model unit is specifically configured to train the first machine learning model established in advance by using the expiration service data, the existing service data and the re-pricing deadline information as training samples, so as to obtain an asset liability configuration model in an initial re-pricing deadline mode and a corresponding training result.

In this embodiment, the configuration model unit is further specifically configured to train the first machine learning model that is built in advance by using the expiration service data, the existing service data, and the remaining term information as training samples, so as to obtain an asset liability configuration model and a corresponding training result in an initial remaining term mode.

In this embodiment, the configuration model unit is further specifically configured to optimize the asset liability configuration model in the initial re-pricing deadline mode and the asset liability configuration model in the initial remaining deadline mode by using the acquired newly-added service data as a test sample, and generate the trained asset liability configuration model in the re-pricing deadline mode and the corresponding training result, and the trained asset liability configuration model in the remaining deadline mode and the corresponding training result.

The pricing model unit is specifically configured to use a training result of the asset liability configuration model in the initial re-pricing deadline mode as a training sample, and obtain the internal funds transfer cost pricing information in the re-pricing deadline mode through training of the initial internal funds transfer cost pricing model.

The pricing model unit is further specifically configured to use a training result of the asset liability configuration model in the initial remaining period mode as a training sample, and obtain the internal funds transfer cost pricing information in the remaining period mode through initial internal funds transfer cost pricing training of the model.

The contribution prediction unit is specifically configured to obtain, through calculation of a rolling differential, service performance contribution information in the re-pricing deadline mode according to the pricing information of the internal funds transfer cost in the re-pricing deadline mode.

The contribution prediction unit is further specifically configured to obtain service performance contribution information in the remaining deadline mode through calculation according to the pricing information of the internal funds transfer cost in the remaining deadline mode.

Based on the same application conception as the banking performance contribution information prediction method, the invention also provides the banking performance contribution information prediction device. The principle of solving the problem of the banking performance contribution information prediction device is similar to that of a banking performance contribution information prediction method, so that the implementation of the banking performance contribution information prediction device can refer to the implementation of the banking performance contribution information prediction method, and repeated parts are omitted.

Fig. 4 is a schematic structural diagram of a banking performance contribution information prediction system according to an embodiment of the present invention, where the system includes: the system comprises a data acquisition module 1, an asset liability configuration training module 2, an internal funds transfer cost pricing training module 3, a business performance contribution prediction module 4 and a data presentation module 5.

As shown in fig. 5, the banking data acquisition module structure is composed of a source data loading unit 11, a source data preprocessing unit 12 and a data transmission unit 13. The method has the function of loading, processing and transmitting data generated by various business systems of the bank.

Source data loading unit 11: the characteristic value data is provided to the equity configuration training module 2 by receiving and loading business data generated by various types of application servers related to the equity liabilities of the bank, such as a selection period of historical business data, a selection business system (deposit system, loan system, trade financing system, financial market system, clearing system, etc.).

Source data preprocessing unit 12: classifying the business data to obtain asset business data and liability business data; screening service data, and extracting characteristic information such as asset liability information, customer basic information, service product information, currency information, principal information, loss benefit information, deadline information (date of interest, due date, and re-pricing date); processing special data, including:

the system missing data is subjected to supplementary recording, and key characteristic information of the system missing is supplemented; the information which is not supported to be input by the system is subjected to the supplementary recording, for example, the rights and interests of owners remained after the liabilities are matched with the assets, the service data details of the system often have no rights and interests of the owners, and an open interface is needed to provide a key information supplementary recording function at the moment;

And processing special rule data, such as special scenes of loan selling, negative interest rate and the like, and judging and processing.

A data transmission unit 13: the preprocessed data is transmitted to the asset liability configuration training module 2 for training.

As shown in fig. 6, a block diagram of the asset liability configuration training module 2 in this embodiment includes: a characteristic value data import 21, a machine learning unit 22, and a result data transmission unit 23. The asset liability configuration training module 2 trains through machine learning based on the characteristic value data and the history result data, and continuously improves the machine learning model through optimization of the characteristic value.

The feature value data import 21 includes: the feature values are imported, and the feature values in this embodiment include asset liability information, customer information (customer number, customer name), business product information (business product category, asset liability identification), currency information (business currency, interest currency), principal information (business principal), profit-loss information (business interest rate, business interest), term information (date of interest, due date, date of re-pricing, remaining term, re-pricing term) and provide a data import for a long period of time. The characteristic value data preprocessing comprises the following steps: assets and liability deadlines are divided and categorized.

Machine learning unit 22: and outputting the result information of each leaf node according to the characteristic value data generated by processing and the judging conditions through a machine learning decision tree algorithm.

The characteristic value data generated by the input processing includes: business currency, asset liability identification, categorized asset deadline codes and liability deadline codes, summarized business principal and business interest. Decision tree decision conditions and decision results refer to fig. 2.

The model training result information comprises: the business currencies under the re-pricing deadline mode, the asset deadline codes correspond to the configuration amount of the liability deadline codes and the owner rights and interests information; and under the residual period mode, each business currency, each asset period code corresponds to the configuration amount of each liability period code and the owner rights and interests information. Wherein, the configuration amount is explained in detail as: and if the liability amount of the term N is greater than or equal to the remaining amount of the asset, assigning the remaining amount of the asset, and if the liability amount is less than the remaining amount of the asset, assigning the liability amount. Period M asset remaining amount = period M asset amount-all configured amounts between liability amounts less than liability period N and period M asset amount. When the term N is 1, the term M asset remaining amount=the term M asset amount. The configured amount may specify a period of the liability amount and a period of the asset amount at the same time. For example, if the amount of liabilities with a period of 1M is 100 tens of thousands, the amount of liabilities with a period of 3M is 100 tens of thousands, the amount of assets with a period of 1M is 150 tens of thousands, and the amount of assets with a period of 3M is 150 tens of thousands, the amount of liabilities with a period of 1M and the amount of assets with a period of 1M are 100 tens of thousands; the amount of configuration of the period 3M liability and the period 1M asset is 50 ten thousand; the amount of configuration of the period 3M liabilities and period 3M assets is 50 tens of thousands.

The result data transmission unit 23: and transmitting the training result to an internal funds transfer cost pricing training model, and inputting the training result as characteristic value data. And transmitting the training result to a data display module to display the configuration condition of the liability.

As shown in fig. 7, the internal funds-transfer cost pricing training model 3 includes: a characteristic value data importing unit 31, a machine learning unit 32, and a result data transmitting unit 33.

The characteristic value data importing unit 31 includes importing training samples and characteristic value data as the internal funds transfer cost pricing training model 3 according to the training results of the asset liability configuration training module 2.

Machine learning unit 32: by multiple linear regression supporting the vector machine regression (SVR) algorithm, by giving a new input sample x, it is deduced from the given data sample what its corresponding output Y is, which is a real number. Regression problems can be described in mathematical language as:

The result data transmission unit 33 transmits the training result to the business performance contribution prediction module 4 as a calculation input parameter of the business performance contribution prediction module. And transmitting the training result to a data display module to display the internal funds transfer cost pricing condition.

As shown in fig. 8, the business performance contribution prediction module 4 includes: a data importing unit 41, a model calculating unit 42 and a result data transmitting unit 43.

The data importing unit 41 includes, as a calculation parameter, an importing business performance contribution prediction module 4 according to the training result of the internal funds transfer cost pricing training module 3; and according to the data acquired by the data acquisition module 1, the data is used as basic data to be imported into the business performance contribution prediction module 4.

The model calculation unit 42 includes, according to the internal funds transfer cost pricing of each term obtained by training by the internal funds transfer cost pricing training module 3, the basic data acquired by the data acquisition module 1, and obtains the performance contribution value of each business by using the rolling algorithm.

The specific rolling algorithm is as follows:

The result data transmission unit 43 includes outputting the service performance contribution value to the data presentation module 5 to present the service performance contribution prediction result.

As shown in fig. 9, the data presentation module 5 includes: a data importing unit 51, an asset liability configuration data presentation unit 52, an internal funds transfer cost pricing data presentation unit 53, and a business performance contribution prediction data presentation unit 54.

The data importing unit 51 includes importing result data of the asset liability configuration training module 2, the internal funds transfer cost pricing training module 3, and the business performance contribution prediction module 4.

The balance configuration data presentation unit 52 presents the training result of the balance configuration training module 2.

The internal funds transfer cost pricing data presentation unit 53 presents the training results of the internal funds transfer cost pricing training module 3.

The business performance contribution prediction data presentation unit 54 presents the prediction result of the business performance contribution prediction module 4.

In this embodiment, the eigenvalue source data: including data for individual business categories associated with liabilities. Since the business categories involved in each commercial bank are different, as shown in tables 4 to 7, it includes:

table 4A class of service data

Table 5B class of service data

Table 6C class of service data

Table 7D class traffic data:

class a services include: deposit and loan service information (customer number, customer name, service currency, service principal, service interest, date of interest, expiration date, remaining term, and re-pricing term)

Class B services include: trade financing system business information (customer number, customer name, business currency, business principal, business interest, day of interest, due date, remaining period, re-pricing period);

class C services include: financial market business information (customer number, customer name, business currency, business principal, business interest, date of interest, expiration date, remaining term, re-pricing term);

Class D services include: system business information (customer number, customer name, business currency, business principal, business interest) is cleared.

The asset liability configuration training module trains result data: and generating a bank asset liability configuration situation through model training according to the expiration service data, the existing service data, the asset data, the liability data and the characteristic value information of the banking service application system.

The asset liability configuration training result data structure in this embodiment is shown in table 8.

TABLE 8

The internal funds transfer cost pricing training module trains the outcome data: and generating bank internal funds transfer cost pricing through model training according to the result data of the asset liability configuration training module.

The internal funds-transfer cost pricing training results data structures in this embodiment are shown in Table 9.

TABLE 9

Business performance contribution prediction outcome data: and generating a service performance contribution predicted value according to the internal funds transfer cost pricing training module result data.

The banking performance contribution prediction data structure in this embodiment is shown in table 10.

Table 10

As shown in fig. 10, the electronic device 600 may further include: a communication module 110, an input unit 120, an audio processing unit 130, a display 160, a power supply 170. It is noted that the electronic device 600 need not include all of the components shown in fig. 10; in addition, the electronic device 600 may further include components not shown in fig. 10, to which reference is made to the related art.

As shown in fig. 10, the central processor 100, sometimes also referred to as a controller or operational control, may include a microprocessor or other processor device and/or logic device, which central processor 100 receives inputs and controls the operation of the various components of the electronic device 600.

The memory 140 may be, for example, one or more of a buffer, a flash memory, a hard drive, a removable media, a volatile memory, a non-volatile memory, or other suitable device. The information about failure may be stored, and a program for executing the information may be stored. And the central processor 100 can execute the program stored in the memory 140 to realize information storage or processing, etc.

The input unit 120 provides an input to the central processor 100. The input unit 120 is, for example, a key or a touch input device. The power supply 170 is used to provide power to the electronic device 600. The display 160 is used for displaying display objects such as images and characters. The display may be, for example, but not limited to, an LCD display.

The memory 140 may be a solid state memory such as Read Only Memory (ROM), random Access Memory (RAM), SIM card, or the like. But also a memory which holds information even when powered down, can be selectively erased and provided with further data, an example of which is sometimes referred to as EPROM or the like. Memory 140 may also be some other type of device. Memory 140 includes a buffer memory 141 (sometimes referred to as a buffer). The memory 140 may include an application/function storage 142, the application/function storage 142 for storing application programs and function programs or a flow for executing operations of the electronic device 600 by the central processor 100.

The memory 140 may also include a data store 143, the data store 143 for storing data, such as contacts, digital data, pictures, sounds, and/or any other data used by the electronic device. The driver storage 144 of the memory 140 may include various drivers of the electronic device for communication functions and/or for performing other functions of the electronic device (e.g., messaging applications, address book applications, etc.).

The communication module 110 is a transmitter/receiver 110 that transmits and receives signals via an antenna 111. A communication module (transmitter/receiver) 110 is coupled to the central processor 100 to provide an input signal and receive an output signal, which may be the same as in the case of a conventional mobile communication terminal.

Based on different communication technologies, a plurality of communication modules 110, such as a cellular network module, a bluetooth module, and/or a wireless local area network module, etc., may be provided in the same electronic device. The communication module (transmitter/receiver) 110 is also coupled to a speaker 131 and a microphone 132 via an audio processor 130 to provide audio output via the speaker 131 and to receive audio input from the microphone 132 to implement usual telecommunication functions. The audio processor 130 may include any suitable buffers, decoders, amplifiers and so forth. In addition, the audio processor 130 is also coupled to the central processor 100 so that sound can be recorded locally through the microphone 132 and so that sound stored locally can be played through the speaker 131.

It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as a method, system, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present invention is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

The principles and embodiments of the present invention have been described in detail with reference to specific examples, which are provided to facilitate understanding of the method and core ideas of the present invention; meanwhile, as those skilled in the art will have variations in the specific embodiments and application scope in accordance with the ideas of the present invention, the present description should not be construed as limiting the present invention in view of the above.

Claims

1. A banking performance contribution information prediction method, the method comprising:

Training a first machine learning model established in advance by taking the expiration service data, the existing service data and the characteristic information as training samples to obtain an initial asset liability configuration model and a corresponding training result; the training result of the asset liability configuration model comprises the following steps: the business currencies under the re-pricing deadline mode, the asset deadline codes correspond to the configuration amount of the liability deadline codes and the owner rights and interests information; each business currency under the residual period mode, and each asset period code corresponds to the configuration amount of each liability period code and owner rights and interests information;

2. The method of claim 1, wherein the data processing the service data comprises: and carrying out supplementary recording on the missing service data and processing special rule data in the service data.

3. The method of claim 1, wherein the performing data processing on the business data to obtain asset business data and liability business data comprises: and classifying the expired service data and the existing service data to obtain asset service data and liability service data.

4. The method of claim 1, wherein the performing data processing on the service data to obtain feature information corresponding to the asset service data and liability service data further comprises: and screening the expired service data and the existing service data, and extracting characteristic information, wherein the characteristic information comprises customer basic information, service product information, currency information, principal information, damage and benefit information, re-pricing deadline information and residual deadline information.

5. The method of claim 4, wherein training the pre-established first machine learning model using the expiration business data, the existing business data, and the feature information as training samples to obtain an initial asset liability configuration model and corresponding training results comprises: and training a pre-established first machine learning model by taking the expiration service data, the existing service data and the re-pricing deadline information as training samples to obtain an initial asset liability configuration model in a re-pricing deadline mode and a corresponding training result.

6. The method of claim 5, wherein training the pre-established first machine learning model using the expiration business data, the existing business data, and the feature information as training samples to obtain an initial asset liability configuration model and corresponding training results further comprises: and training a first machine learning model which is built in advance by taking the expiration service data, the existing service data and the residual deadline information as training samples to obtain an initial asset liability configuration model under the residual deadline mode and a corresponding training result.

7. The method of claim 6, wherein training the pre-established first machine learning model using the expiration business data, the existing business data, and the feature information as training samples, after obtaining an initial asset liability configuration model and corresponding training results, further comprises: and optimizing the asset liability configuration model in the initial re-pricing deadline mode and the asset liability configuration model in the initial residual deadline mode by taking the acquired newly-added business data as a test sample, and generating an asset liability configuration model in the trained re-pricing deadline mode and a corresponding training result, and an asset liability configuration model in the trained residual deadline mode and a corresponding training result.

8. The method of claim 6, wherein the training results of the initial internal funds-transfer cost pricing model include internal funds-transfer cost pricing information in a re-pricing deadline mode, the internal funds-transfer cost pricing information generated by: and training the initial internal funds transfer cost pricing model by taking the training result of the asset liability configuration model in the initial re-pricing deadline mode as a training sample to obtain the internal funds transfer cost pricing information in the re-pricing deadline mode.

9. The method of claim 6, wherein the training results of the initial internal funds-transfer cost pricing model include internal funds-transfer cost pricing information in a remaining deadline mode, the internal funds-transfer cost pricing information generated by: and training the initial internal funds transfer cost pricing model by taking the training result of the asset liability configuration model in the initial remaining period mode as a training sample to obtain the internal funds transfer cost pricing information in the remaining period mode.

10. The method of claim 8, wherein calculating business performance contribution information using a rolling algorithm using training results of an initial internal funds transfer cost pricing model as calculation parameters comprises: and obtaining service performance contribution information in the re-pricing deadline mode through rolling calculation according to the internal funds transfer cost pricing information in the re-pricing deadline mode.

11. The method of claim 9, wherein calculating business performance contribution information using a rolling algorithm using training results of an initial internal funds transfer cost pricing model as calculation parameters comprises: and obtaining service performance contribution information in the remaining period mode through rolling calculation according to the internal funds transfer cost pricing information in the remaining period mode.

12. A banking performance contribution information prediction apparatus, the apparatus comprising:

the configuration model unit is used for training a first machine learning model which is built in advance by taking the expiration service data, the existing service data and the characteristic information as training samples to obtain an initial asset liability configuration model and a corresponding training result; the training result of the asset liability configuration model comprises the following steps: the business currencies under the re-pricing deadline mode, the asset deadline codes correspond to the configuration amount of the liability deadline codes and the owner rights and interests information; each business currency under the residual period mode, and each asset period code corresponds to the configuration amount of each liability period code and owner rights and interests information;

13. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the processor implements the method of any one of claims 1 to 11 when executing the program.

14. A computer readable storage medium, characterized in that the computer readable storage medium stores a computer program for executing the method of any one of claims 1 to 11.