CN116882550A

CN116882550A - Balance change prediction method, system and computer equipment

Info

Publication number: CN116882550A
Application number: CN202310717835.8A
Authority: CN
Inventors: 占健智
Original assignee: Nanjing Xingyun Digital Technology Co Ltd
Current assignee: Nanjing Xingyun Digital Technology Co Ltd
Priority date: 2023-06-16
Filing date: 2023-06-16
Publication date: 2023-10-13

Abstract

The application discloses a balance change prediction method, a system and computer equipment, wherein the method comprises the following steps: screening historical large-amount transaction account data based on preset rules and a historical data set, wherein the large-amount transaction data comprises a large-amount transaction account and corresponding deposit amount data thereof; acquiring a deposit influencing index based on deposit amount data corresponding to the historical large transaction account; establishing and training a prediction model based on the stock effect index and the historical dataset; carrying out deposit withdrawal prediction on the current account based on the prediction model; the overall balance variation can be effectively controlled; and refining the deposit influence index, and establishing a training prediction model according to the deposit influence index, so that the deposit behavior of the customer can be predicted in a classified manner, and the actions of early warning deposit and predicting deposit can be achieved.

Description

Balance change prediction method, system and computer equipment

Technical Field

The application relates to the field of data processing, in particular to a balance change prediction method, a balance change prediction system and computer equipment.

Background

In order to ensure stable operation of the financial platform, the balance of the commercial bank needs to be stably increased, so that analysis of the variation of the balance of the commercial bank is extremely important. Increasing balance conversion requires accurate analysis of customers with primary balance variation and differentiated management of these customers, thereby reducing large withdrawals and increasing/maintaining large deposits.

However, at present, it is difficult for the financial platform to accurately identify the customers with main balance variation, and it is also difficult to predict the deposit behavior of the customers with large deposit and withdrawal, so that the effect of estimating the influence of the customers on the overall balance variation is poor, and the risk management effect of the financial platform commercial bank balance is poor.

Disclosure of Invention

The application aims at: a balance change prediction method, system and computer device are provided.

The technical scheme of the application is as follows: in a first aspect, the present application provides a balance variation prediction method, the method comprising:

screening historical large-amount transaction account data based on preset rules and a historical data set, wherein the large-amount transaction data comprises a large-amount transaction account and corresponding deposit amount data thereof;

acquiring a deposit influencing index based on deposit amount data corresponding to the historical large transaction account;

establishing and training a prediction model based on the stock effect index and the historical dataset;

and carrying out deposit withdrawal prediction on the current account based on the prediction model.

In a preferred embodiment, the screening out historical large transaction account data based on the preset rules and the historical data set includes:

calculating a median of annual average withdrawal according to the historical dataset;

and screening historical large transaction account data with the month deposit amount larger than the preset difference value of the number of average deposit withdrawal digits or larger than the preset multiple of the number of average deposit withdrawal digits.

In a preferred embodiment, the obtaining the deposit influence indicator based on the deposit amount data corresponding to the historical large transaction account includes:

analyzing the historical large transaction account data to obtain related storage effect indexes;

and screening the related promotion effect indexes based on a logistic regression model to obtain target promotion effect indexes.

In a preferred embodiment, the screening the relevant promotion effect indicators based on the logistic regression model to obtain the target promotion effect indicators includes:

dividing the historical large transaction account data according to the age dimension;

analyzing historical large transaction account data of each age dimension based on a logistic regression model and the related lifting impact index to obtain an analysis result;

and screening the relevant deposit influence indexes based on the analysis result to obtain target deposit influence indexes.

In a preferred embodiment, the analyzing the historical large transaction account data for each of the age dimensions based on the logistic regression model and the associated inventory impact indicators to obtain the analysis results includes:

performing logistic regression model analysis on historical large transaction account data of each age dimension by using the related promotion effect indexes to obtain B coefficients of each related promotion effect index in each age dimension;

the screening the relevant deposit influence indexes based on the analysis result to obtain target deposit influence indexes comprises the following steps:

and screening the relevant deposit influence indexes with the B coefficient larger than 0 as target deposit influence indexes.

In a preferred embodiment, said building and training a predictive model based on said load-bearing impact indicator and said historical dataset comprises:

constructing a behavior prediction model based on the target storage influence indexes by utilizing a decision tree algorithm;

training the behavior prediction model based on the historical dataset to obtain a prediction model.

In a preferred embodiment, before the building and training of the predictive model based on the storage impact indicator and the historical dataset, the method further comprises:

performing cluster analysis on the historical large transaction account data to obtain a cluster classification result;

after the training the behavior prediction model based on the historical dataset to obtain a prediction model, the method further comprises:

verifying the behavior prediction model based on the clustering classification result;

and if the verification is passed, obtaining a prediction model.

In a preferred embodiment, the predicting the withdrawal of the current account based on the prediction model includes:

obtaining target high deposit customer data based on the predictive model and current account data;

and obtaining a predicted large-amount withdrawal client list based on the prediction model and the preset large-amount withdrawal risk early warning threshold signal and the target large-amount deposit client data.

In a second aspect, the present application also provides a balance change prediction system, the system comprising:

the screening module is used for screening historical large-amount transaction account data based on preset rules and a historical data set, wherein the large-amount transaction data comprise large-amount transaction accounts and corresponding withdrawal amount data;

the acquisition module is used for acquiring a deposit influence index based on deposit amount data corresponding to the historical large transaction account;

the training module is used for building and training a prediction model based on the storage influence index and the historical data set;

and the prediction module is used for predicting the withdrawal of the current account based on the prediction model.

In a third aspect, the present application also provides a computer apparatus comprising:

one or more processors;

and a memory associated with the one or more processors, the memory for storing program instructions that, when read for execution by the one or more processors, perform the balance variation prediction method of any of the first aspects.

The application has the advantages that: provided are a balance change prediction method, a balance change prediction system and computer equipment, wherein the balance change prediction method comprises the following steps: screening historical large-amount transaction account data based on preset rules and a historical data set, wherein the large-amount transaction data comprises a large-amount transaction account and corresponding deposit amount data thereof; acquiring a deposit influencing index based on deposit amount data corresponding to the historical large transaction account; establishing and training a prediction model based on the stock effect index and the historical dataset; carrying out deposit withdrawal prediction on the current account based on the prediction model; the overall balance variation can be effectively controlled; and refining the deposit influence index, and establishing a training prediction model according to the deposit influence index, so that the deposit behavior of the customer can be predicted in a classified manner, and the actions of early warning deposit and predicting deposit can be achieved.

Drawings

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

FIG. 1 is a flow chart of a balance variation prediction method provided by the application;

FIG. 2 is a schematic diagram showing a logistic regression model for each age in the balance change prediction method according to the present application analyzing the group by using the four related lifting effect indexes and the analysis result graph of the large lifting;

FIG. 3 is a schematic diagram of a prediction model in the balance variation prediction method provided by the application;

FIG. 4 is a diagram showing the result of transforming a decision tree into a classification list in the balance change prediction method according to the present application;

FIG. 5 is a schematic diagram of model application and marketing strategy adaptation in the balance variation prediction method provided by the application;

FIG. 6 is a block diagram of a balance change prediction system according to the present application;

fig. 7 is a schematic diagram of a computer device according to the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are only some embodiments of the present application, not all embodiments of the present application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.

As described in the background art, in order to ensure the stability of the financial platform, in the prior art, the balance of the financial platform needs to be controlled, but the balance deposit on the financial platform is carried out by different customers, so that it is necessary to find out customers who carry out large amounts of deposit, and predict the deposit withdrawal behaviors of the customers, so that means such as marketing can be pertinently executed, thereby stabilizing the deposit withdrawal behaviors of the customers and ensuring the balance stability of the financial platform.

In order to solve the problems, the application creatively provides a balance change prediction method, a balance change prediction system and computer equipment, which can accurately predict the large-amount lifting of customers and ensure the balance stability of a financial platform.

The following describes the embodiments of the present application in detail with reference to the drawings.

Embodiment one: this embodiment describes a balance fluctuation predicting process according to the present application with reference to fig. 1.

Specifically, the present embodiment provides a balance variation prediction method, which includes:

s110, screening historical large-amount transaction account data based on preset rules and a historical data set, wherein the large-amount transaction account data comprise large-amount transaction accounts and corresponding deposit amount data.

In one embodiment, the screening out historical large transaction account data based on the preset rules and the historical dataset includes:

and S111, calculating the median of annual average withdrawal according to the historical data set.

Illustratively, the historical data set contains 10 ten thousand customers' historical data including deposit amount, deposit time, withdrawal amount and withdrawal time, customer background information, etc. for each customer. Customers with large transactions, whether withdrawal or deposit, account for only 20% of the total customers, but their transaction amount accounts for nearly 80% of the total transaction amount. Most importantly, their transaction amount is 3-4 times greater than non-high volume transaction customers, and their high volume transactions account for nearly 35% of all customers' transaction amounts. So a large change in balance is coming from this large group of traders. The median of deposit was calculated.

And S112, screening out historical large transaction account data with the month deposit amount larger than the preset difference value of the average deposit amount or larger than the preset multiple of the average deposit amount.

Illustratively, historical large transaction account data for monthly offers, i.e., single month offers or deposits, are screened for an amount greater than 1.5 times the median annual average deposit, or for a monthly offer greater than the median annual average deposit rmb$1000.

S120, acquiring a deposit influence index based on deposit amount data corresponding to the historical large-amount transaction account.

In one embodiment, the obtaining the deposit influence index based on the deposit amount data corresponding to the historical large transaction account includes:

s121, analyzing the historical large transaction account data to obtain relevant withdrawal influence indexes.

Specifically, four relevant deposit influencing indexes are obtained based on analysis of the withdrawal amount, the withdrawal time, the deposit amount and the deposit time of the historical large transaction account data: the volatility of the promotion (number of times of large promotion in common) -representing the volatility of the customer's promotion-may be related to his consumption/investment pattern; aggression (total withdrawal/total deposit ratio) -representing whether the customer is of the aggressive or conservative type; withdrawal ranking (total withdrawal/average withdrawal for the same age group) -represents whether the customer belongs to a group of more than withdrawal in the age group; deposit ranking (total deposit amount/average deposit amount for the same age group) -represents whether the customer belongs to a group of more than deposit in the age group.

S122, screening the relevant deposit influence indexes based on a logistic regression model to obtain target deposit influence indexes.

In one embodiment, the step comprises:

s1221, dividing the historical large transaction account data according to the age dimension.

S1222, analyzing historical large transaction account data of each age dimension based on the logistic regression model and the related storage influence indexes to obtain an analysis result.

Each associated deposit impact indicator estimates that they will be associated with an amount that will cause the customer to make a large deposit during the month described above. For example, it is estimated that persons who have a higher amount to be deposited than the same age, and who are aggressive, should have a greater deposited amount. Thus, a logistic regression model analysis was performed (logistic regression model). Logistic regression model analysis was performed separately for each age group with the above four relevant load-up impact indicators and the maximum load-up, and the analysis results are shown in fig. 2.

Preferably, the step includes:

and performing logistic regression model analysis on the historical large transaction account data of each age dimension by using the relevant promotion effect index to obtain a B coefficient of each relevant promotion effect index in each age dimension.

S1223, screening the relevant deposit influence indexes based on the analysis result to obtain target deposit influence indexes.

Preferably, this step:

Specifically, the larger the B-factor, the more characteristic the relevant promotion impact index is representative of a customer possessing a large promotion. Within all age groups, the B-factor for withdrawal and deposit ranks is zero, meaning that customers with large withdrawal will be present either with high withdrawal or low withdrawal, rather than only with high withdrawal (extended to high net asset value).

On the contrary, the B coefficient of the stock volatility and the aggressiveness is very remarkable, which proves that the two factors can be important indexes for distinguishing the stock with large amount and the stock with non-large amount.

Thus, the target deposit influence indexes are screened out, including deposit volatility and deposit aggressiveness.

SA10, carrying out cluster analysis on the historical large transaction account data to obtain a cluster classification result.

Each customer will have his own specific stock keeping law due to his own background. In the event that two customers have similar backgrounds (e.g., age, income, home background, occupation, etc.), they have a close law of large inventory. Based on the present only per month inventory data, in other words, the same class of customers will have a large inventory transaction in the next month. A clustering algorithm is used to cluster clients with similar backgrounds and large transactions in the same month.

S130, building and training a prediction model based on the storage influence index and the historical data set.

The prediction model is constructed by the idea shown in figure 3.

In one embodiment, the building and training of the predictive model based on the inventory impact indicators and the historical dataset includes:

s131, constructing a behavior prediction model based on the target storage influence index by utilizing a decision tree algorithm.

And S132, training the behavior prediction model based on the historical data set to obtain a prediction model.

Specifically, using Python software, a model of running behavior prediction is used to build customer classification criteria for each age group using a decision tree algorithm. The following is a definition of model classification:

decision trees for data mining are implemented with Python. The transformation of the decision tree into a sorted list (classification table) is shown in fig. 4. Wherein (1) the region represents customers meeting the fluctuation and access of the promotion are marked as specific large-scale customers, and (2) the region represents specific non-large-scale promotion customers; while the numbers in the table represent the true proportion (%) of the particular large proposed user of the classification. Taking the example of fig. 4, if a customer has a volatility of 3 and an accessibility of 0.95, he is classified as a specific high-volume inventory (high-volume inventory in 8-10 months, which accounts for about 70% of the total inventory), and a group that needs to be managed/controlled. And the accuracy of this classification is 85.2%.

From the sorted list, it is seen that a particular large percentage of the inventory (customers with large percentage of the inventory in a particular month, but not in other months) tends to have a larger number of large percentage of the inventory and percentage of the inventory. This is also a reasonable differentiation result, indirectly justifying the model.

SA20, verifying the behavior prediction model based on the clustering classification result;

and if the verification is passed, obtaining a prediction model.

To verify the accuracy of the above classification model, the following analysis (fusion Matrix) was performed:

at present, identification is carried out based on two target promotion influence indexes of promotion volatility and aggressiveness, the classification accuracy of the model reaches more than 75 percent-!

S140, carrying out deposit withdrawal prediction on the current account based on the prediction model.

In one embodiment, the predicting the withdrawal for the current account based on the predictive model includes:

s141, obtaining target large deposit client data based on the prediction model and the current account data.

And S142, obtaining a predicted large-amount withdrawal client list based on the prediction model and the preset large-amount withdrawal risk early warning threshold signal and the target large-amount deposit client data.

And setting a large-amount extraction risk early warning threshold signal and a deposit marketing signal, and performing policy adaptation on a case-by-case basis according to different guest groups, and providing the marketing personnel with early intervention for guiding clients, as shown in fig. 5.

Marketing is contacted to clients on the predicted large-amount withdrawal client list. Based on the portrait, the label and the different touch modes of the whole channel of the forecast large-amount money drawing client, operators can formulate a marketing strategy of single event or integrated event, and support the automatic marketing of the event and the wave number so as to achieve the accurate marketing to the client.

Embodiment III: corresponding to the above-described embodiments, a balance change prediction system according to the present application will be described with reference to fig. 6. The system may be implemented in hardware or software, or may be implemented in a combination of hardware and software, which is not limited by the present application.

In one example, the present application provides a balance change prediction system comprising:

a screening module 610, configured to screen out historical large transaction account data based on a preset rule and a historical dataset, where the large transaction data includes a large transaction account and corresponding withdrawal amount data thereof;

an obtaining module 620, configured to obtain a deposit influencing indicator based on deposit amount data corresponding to the historical large transaction account;

a build training module 630 for building and training a predictive model based on the stock effect index and the historical dataset;

and the prediction module 640 is used for predicting the withdrawal of the current account based on the prediction model.

In one embodiment, the screening module 610 includes:

a calculating unit 611 for calculating a median of annual average withdrawal from the history data set;

a first screening unit 612 is configured to screen historical large transaction account data with a month deposit amount greater than the preset difference of the number of average deposit withdrawal digits or greater than the preset multiple of the number of average deposit withdrawal digits.

Preferably, the acquiring module 620 includes:

an analysis unit 621, configured to analyze the historical large transaction account data to obtain relevant withdrawal impact indicators;

a second screening unit 622, configured to screen the relevant promotion impact indicators based on a logistic regression model to obtain target promotion impact indicators.

More preferably, the second screening unit 622 includes:

a dividing sub-unit 6221 for dividing the historical large transaction account data according to the age dimension;

an analysis subunit 6222, configured to analyze the historical large transaction account data of each age dimension based on a logistic regression model and the relevant withdrawal impact indicators to obtain an analysis result;

and a screening subunit 6223, configured to screen the relevant promotion impact indicators based on the analysis result to obtain target promotion impact indicators.

More preferably, the analysis subunit 6222 is specifically configured to:

More preferably, the building training module 630 includes:

a construction unit 631 for constructing a behavior prediction model based on the target storage influence index by using a decision tree algorithm;

a training unit 632 is configured to train the behavior prediction model to obtain a prediction model based on the historical dataset.

More preferably, the system further comprises:

the clustering module 650 is configured to perform cluster analysis on the historical large transaction account data to obtain a cluster classification result before the building training module 630 builds and trains a prediction model based on the storage influence index and the historical data set;

the system further comprises:

a verification module 660 for verifying the behavior prediction model based on the cluster classification result after the build training module 630 builds and trains a prediction model based on the presence impact index and the historical dataset;

if the verification module 660 verifies, a prediction model is obtained.

More preferably, the prediction module 640 includes:

a first obtaining unit 641 for obtaining target high deposit customer data based on the prediction model and current account data;

a second obtaining unit 642, configured to obtain a list of predicted large-amount withdrawal customers based on the prediction model and a preset large-amount withdrawal risk early-warning threshold signal and the target large-amount deposit customer data.

Embodiment four: corresponding to the first to third embodiments, the computer device provided by the present application will be described with reference to fig. 7. In one example, as shown in FIG. 7, the present application provides a computer device comprising:

one or more processors;

and a memory associated with the one or more processors, the memory for storing program instructions that, when read for execution by the one or more processors, perform the following:

The program instructions, when read for execution by the one or more processors, further perform the operations of:

and if the verification is passed, obtaining a prediction model.

The program instructions, when read and executed by the one or more processors, may further perform operations corresponding to the steps in the foregoing method embodiments, and reference may be made to the foregoing description, which is not repeated herein. With reference to FIG. 7, an exemplary architecture for a computer device is shown, which may include a processor 710, a video display adapter 711, a disk drive 712, an input/output interface 713, a network interface 714, and a memory 720. The processor 710, the video display adapter 711, the disk drive 712, the input/output interface 713, the network interface 714, and the memory 720 may be communicatively connected via a communication bus 730.

The processor 710 may be implemented by a general-purpose central processing unit (Central Processing Unit, CPU), a microprocessor, an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., for executing related programs to implement the technical scheme provided by the present application.

The Memory 720 may be implemented in the form of Read Only Memory (ROM), random access Memory (Random Access Memory, RAM), static storage devices, dynamic storage devices, and the like. The memory 720 may store an operating system 721 for controlling the operation of the computer device 700, and a Basic Input Output System (BIOS) 722 for controlling the low-level operation of the computer device 700. In addition, a web browser 723, data storage management 724, and icon font processing system 725, etc. may also be stored. The icon font processing system 725 may be an application program that specifically implements the operations of the foregoing steps in the embodiment of the present application. In general, when the technical solution provided by the present application is implemented by software or firmware, relevant program codes are stored in the memory 720 and invoked by the processor 710 for execution.

The input/output interface 713 is used to connect with an input/output module to enable information input and output. The input/output module may be configured as a component in a device (not shown) or may be external to the device to provide corresponding functionality. Wherein the input devices may include a keyboard, mouse, touch screen, microphone, various types of sensors, etc., and the output devices may include a display, speaker, vibrator, indicator lights, etc.

The network interface 714 is used to connect communication modules (not shown) to enable communication interactions of the device with other devices. The communication module may implement communication through a wired manner (such as USB, network cable, etc.), or may implement communication through a wireless manner (such as mobile network, WIFI, bluetooth, etc.).

Bus 730 includes a path to transfer information between various components of the device (e.g., processor 710, video display adapter 711, disk drive 712, input/output interface 713, network interface 714, and memory 720).

In addition, the computer device 700 may also obtain information of specific retrieval conditions from the virtual resource object retrieval condition information database 741 for making condition judgment, and the like.

It should be noted that although the above-described computer device 700 illustrates only a processor 710, a video display adapter 711, a disk drive 712, an input/output interface 713, a network interface 714, a memory 720, a bus 730, etc., the computer device may include other components necessary to achieve proper operation in an implementation. Furthermore, it will be appreciated by those skilled in the art that the apparatus may include only the components necessary to implement the present application, and not all of the components shown in the drawings.

From the above description of embodiments, it will be apparent to those skilled in the art that the present application may be implemented in software plus a necessary general hardware platform. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions for causing a computer device (which may be a personal computer, a cloud server, or a network device, etc.) to execute the method described in the embodiments or some parts of the embodiments of the present application.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for the device embodiments, since they are substantially similar to the method embodiments, the description is relatively simple, and reference is made to the description of the method embodiments for relevant points. The apparatus embodiments described above are merely illustrative, wherein the modules illustrated as separate components may or may not be physically separate, and the components shown as modules may or may not be physical, i.e., may be located in one place, or may be distributed over a plurality of network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present application without undue burden.

In addition, it is to be understood that: the terms "first" and "second" are used herein for descriptive purposes only and are not to be construed as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defining "a first" or "a second" may explicitly or implicitly include one or more such feature.

The above embodiments are merely for illustrating the technical concept and features of the present application, and are not intended to limit the scope of the present application to those skilled in the art to understand the present application and implement the same. All modifications made according to the spirit of the main technical proposal of the application should be covered in the protection scope of the application.

Claims

1. A balance change prediction method, the method comprising:

2. The balance variation prediction method according to claim 1, wherein the screening out historical large transaction account data based on a preset rule and a historical dataset comprises:

3. The balance change prediction method according to claim 2, wherein the obtaining a deposit influence index based on deposit amount data corresponding to the historical large transaction account includes:

4. The balance variation prediction method according to claim 3, wherein the screening the relevant sequestration influence indicators based on a logistic regression model to obtain target sequestration influence indicators comprises:

5. The balance variation prediction method according to claim 4, wherein analyzing the historical large transaction account data for each of the age dimensions based on the logistic regression model and the associated presence impact indicator to obtain an analysis result comprises:

6. The balance variation prediction method of claim 5, wherein the building and training a prediction model based on the sequestration impact indicator and the historical dataset comprises:

7. The balance variation prediction method of claim 6, wherein prior to the establishing and training of the predictive model based on the sequestration impact indicator and the historical dataset, the method further comprises:

and if the verification is passed, obtaining a prediction model.

8. The balance variation prediction method of claim 6, wherein the predicting the withdrawal for the current account based on the prediction model comprises:

9. A balance variation prediction system, the system comprising:

10. A computer device, the computer device comprising:

one or more processors;

and a memory associated with the one or more processors, the memory for storing program instructions that, when read for execution by the one or more processors, perform the balance variation prediction method of any of claims 1-8.