CN114820038A

CN114820038A - User loss prediction method, device, equipment and medium

Info

Publication number: CN114820038A
Application number: CN202210335042.5A
Authority: CN
Inventors: 黄开旭
Original assignee: Agricultural Bank of China
Current assignee: Agricultural Bank of China
Priority date: 2022-03-31
Filing date: 2022-03-31
Publication date: 2022-07-29

Abstract

The application provides a user loss prediction method. According to the method, user data including time information is obtained through electronic equipment, then the user data is input into a user loss prediction model, and the user data is predicted through a gate control cycle unit and an attention mechanism module of the user loss prediction model to obtain a user loss prediction result. Therefore, the user loss can be predicted through the user data comprising the time information, and the user data comprises the time information, so that the prediction can be performed based on the time characteristics, and the accuracy of the model is improved. In addition, the user loss prediction model introduces an attention mechanism, so that the importance degree of key features can be highlighted, and the accuracy of user loss prediction is further improved.

Description

User loss prediction method, device, equipment and medium

Technical Field

The present application relates to the field of computer technologies, and in particular, to a method, an apparatus, a device, and a computer-readable storage medium for predicting user churn.

Background

With the continuous development of information technology, users are important resources in various applications. With the continuous improvement of the attention of users to investment and financing, many users change the fund originally stored in the bank to buy financing products, thereby causing the loss of the users of the bank.

For banks, it is important to predict the loss of users. By predicting potential lost customers in advance, marketing strategies can be customized for the potential lost customers, or business adjustment can be performed according to the loss trend of the users, so that the loss caused by the loss of the users is avoided.

Therefore, a method for predicting user churn is needed.

Disclosure of Invention

The application provides a user loss prediction method. The method can realize accurate prediction of user loss. The application also provides a device, equipment and a medium corresponding to the method.

In a first aspect, the present application provides a user churn prediction method, including:

acquiring user data including time information;

and inputting the user data into a user loss prediction model, and predicting the user data through a gate control cycle unit and an attention mechanism module of the user loss prediction model to obtain a user loss prediction result.

In some possible implementations, the user churn prediction model is trained by:

acquiring user training data including time information and user loss historical data;

predicting the user training data through a gate control cycle unit and an attention mechanism module of the user loss prediction model to obtain a user loss training prediction result;

and updating parameters of the user loss prediction model according to the user loss training prediction result and the user loss historical data.

In some possible implementations, the user data including the time information includes: and presetting all transaction behaviors of the user and corresponding time in the time.

In some possible implementations, the user data further includes: basic information data of the user.

In some possible implementations, the inputting the user data into a user churn prediction model, and predicting the user data through a gated loop unit and an attention mechanism module of the user churn prediction model to obtain a user churn prediction result includes:

inputting the user data into a user loss prediction model, obtaining a first output through a gate control cycle unit of the user loss prediction model, obtaining attention probability distribution through an attention mechanism module by the first output, processing the attention probability distribution to obtain a second output, inputting the second output into a full connection layer to obtain a third output, and predicting the user data through an activation function by the third output to obtain a user loss prediction result.

In some possible implementations, the attention mechanism module is located in a hidden layer of the user churn prediction model.

In some possible implementations, the method further includes:

and customizing the targeted service according to the user loss prediction result.

In a second aspect, the present application provides a user churn prediction apparatus, which includes:

a communication module for acquiring user data including time information;

and the prediction module is used for inputting the user data into a user loss prediction model, predicting the user data through a gating cycle unit and an attention mechanism module of the user loss prediction model, and obtaining a user loss prediction result.

In some possible implementations, the user churn prediction model is obtained by a user churn prediction model training device configured to:

In some possible implementations, the prediction module is specifically configured to:

In some possible implementations, the apparatus further includes:

and the customizing module is used for customizing the targeted service according to the user loss prediction result.

In a third aspect, the present application provides an apparatus comprising a processor and a memory. The processor and the memory communicate with each other. The processor is configured to execute instructions stored in the memory to cause the apparatus to perform a user churn prediction method as in the first aspect or any implementation of the first aspect.

In a fourth aspect, the present application provides a computer-readable storage medium, where instructions are stored in the computer-readable storage medium, and the instructions instruct a device to perform the user churn prediction method according to the first aspect or any implementation manner of the first aspect.

In a fifth aspect, the present application provides a computer program product comprising instructions that, when run on a device, cause the device to perform the method for user churn prediction as described above in the first aspect or any implementation form of the first aspect.

The present application can further combine to provide more implementations on the basis of the implementations provided by the above aspects.

According to the technical scheme, the embodiment of the application has the following advantages:

the application provides a user loss prediction method, which comprises the steps of obtaining user data comprising time information, inputting the user data into a user loss prediction model, and predicting the user data through a gate control cycle unit and an attention mechanism module of the user loss prediction model to obtain a user loss prediction result. Therefore, the user loss can be predicted through the user data comprising the time information, and the user data comprises the time information, so that the prediction can be performed based on the time characteristics, and the accuracy of the model is improved. In addition, the user loss prediction model introduces an attention mechanism, so that the importance degree of key features can be highlighted, and the accuracy of user loss prediction is further improved.

Drawings

In order to more clearly illustrate the technical method of the embodiments of the present application, the drawings needed to be used in the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without inventive labor.

Fig. 1 is a schematic flowchart illustrating a user churn prediction method according to an embodiment of the present disclosure;

fig. 2 is a schematic diagram of a user churn prediction model according to an embodiment of the present disclosure;

fig. 3 is a schematic diagram illustrating a data flow direction in a user churn prediction model according to an embodiment of the present disclosure;

fig. 4 is a schematic flowchart illustrating a training method of a user churn prediction model according to an embodiment of the present disclosure;

fig. 5 is a schematic structural diagram of a user churn prediction apparatus according to an embodiment of the present disclosure.

Detailed Description

The scheme in the embodiments provided in the present application will be described below with reference to the drawings in the present application.

The terms "first" and "second" in the embodiments of the present application are used 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 defined as "first" or "second" may explicitly or implicitly include one or more of that feature.

Some technical terms referred to in the embodiments of the present application will be first described.

User resources are important resources in various services. As users become more concerned about investing in financing, and the process of the internet for users to purchase financing products becomes easier. Many customers will deposit unused money into the bank to purchase financial products, causing customer loss for the bank.

Therefore, it is an important measure of a commercial bank in facing a strong competition to predict potential customers who are lost in advance and prevent the operation crisis caused by the customer loss for the bank.

Under the general condition, most of the traditional methods adopt a retrospective method to solve the problem of customer loss, namely, the user loss condition is determined by comparing the states of customers before and after marketing activities are carried out, so that a corresponding marketing strategy is formulated, and the method is tedious in process, long in time consumption, low in efficiency and high in cost. Also, this approach is more hysteretic.

In view of the above, the present application provides a method for predicting user churn, which may be performed by an electronic device. An electronic device refers to a device having data processing capabilities and may be, for example, a server or a terminal. The terminal includes, but is not limited to, an id system, a smart phone, a tablet computer, a notebook computer, a Personal Digital Assistant (PDA), and the like. The server may be a cloud server, such as a central server in a central cloud computing cluster, or an edge server in an edge cloud computing cluster. Of course, the server may also be a server in a local data center. The local data center refers to a data center directly controlled by a user.

Specifically, the electronic device acquires user data including time information, inputs the user data into a user loss prediction model, and predicts the user data through a gate control cycle unit and an attention mechanism module of the user loss prediction model to obtain a user loss prediction result. Therefore, the user loss can be predicted through the user data comprising the time information, and the user data comprises the time information, so that the prediction can be performed based on the time characteristics, and the accuracy of the model is improved. In addition, the user loss prediction model introduces an attention mechanism, so that the importance degree of key features can be highlighted, and the accuracy of user loss prediction is further improved.

For convenience of understanding, the user churn prediction method provided by the embodiment of the present application is specifically described below with reference to the drawings.

Referring to fig. 1, a flow chart of a user churn prediction method is shown, the method comprising:

s102: the electronic device obtains user data including time information.

The user data including the time information includes all transaction behaviors of the user within a preset time and corresponding time. The user data also includes basic information data of the user.

Specifically, the user data including the time information includes: the basic information data of the user, all transaction behaviors of the user in the preset time and the corresponding time. The basic information data of the user includes the user age, the user occupation, and the like. All transaction behaviors of the user comprise asset data change, financial data change, loan data change, transaction data change and the like. All transaction behaviors of the user in the preset time and the corresponding time comprise: the user asset data change and the corresponding time, the corresponding time for the user to buy financing, the corresponding time for the user to borrow money, the corresponding time for the user to trade and the like. The preset time can be one year, two years or from the beginning of the account opening for the user to the present.

It should be noted that, in this solution, the user data is acquired after the user explicitly agrees to the user data.

S104: the electronic equipment inputs user data into the user loss prediction model, and the user data is predicted through a gate control cycle unit and an attention mechanism module of the user loss prediction model to obtain a user loss prediction result.

The user attrition prediction model may be framed in fig. 2, where the user data includes basic attributes, asset data, financial data, loan data, and transaction data. T1, t2, and tn are the user data after the user data is preprocessed. Inputting the preprocessed user data into a Gate Recovery Unit (GRU) network, and calculating to obtain corresponding outputs h1, h2,. and hn, which are recorded as first outputs.

The gate control circulation unit Network is a variant with a good effect of a Long Short-Term Memory (LSTM) Network, has one less gate structure compared with the LSTM Network, has a simpler structure, and can be used for solving the problem of Long dependence in a Recurrent Neural Network (RNN). Three gate functions are introduced in LSTM: the input gate, the forgetting gate and the output gate are used for controlling the input value, the memory value and the output value. While in the GRU model there are only two gates: an update gate and a reset gate.

In this embodiment, an attention mechanism module may be added to the user churn prediction model, as shown in fig. 3, to highlight the key information of user churn. Specifically, an attention mechanism, namely an attention module, can be introduced at the hidden layer. The first output (hi) is passed through an attention mechanism module to obtain an attention probability distribution. Specifically, at the time i, as shown in formula (1), an attention probability distribution value ei input before the time i is calculated.

ei＝wi*tanh(Wi*hi+bi) (1)

Wi and Wi represent weight coefficient matrixes at the ith moment, and bi represents corresponding offset at the ith moment.

Further, the output yi (second output) of the attention module can be calculated by equation (2):

yi＝∑ai*hi (2)

wherein ai represents a coefficient, which can be calculated by equation (3):

ai＝exp(ei)/∑exp(ei) (3)

and (3) passing the second output yi of the attention module through the full-connection layer fc to obtain a third output, and then calculating the third output by using the activation function to obtain a user loss prediction result.

In some possible implementations, the method further includes:

s106: and the electronic equipment customizes the targeted service according to the user loss prediction result.

Since the user churn prediction results can be obtained through the above steps, the targeted service can be customized to reduce user churn. The user churn prediction result may include a user group about to churn, a probability that a certain user will churn, or a time that the user will churn.

The user churn prediction model may be obtained by training in the manner shown in fig. 4. Referring to fig. 4, a flowchart of a user churn prediction model training method is shown, where the training method includes:

s402: the electronic device obtains user training data including time information and user churn history data.

The user training data including the time information and the user loss historical data are training data of the model. The user training data including the time information may be data including basic information of the user and all transaction behaviors of the user within a preset time and corresponding time. The user churn history data may be data of whether the user churn.

In some possible implementations, whether a user is away may be determined by the user account content. Such as the user account balance being empty, or the user changing account to 0 within a year, etc.

The training data may be divided in a 7:2:1 ratio, 7/10 may be a training set, 2/10 may be a validation set, and 1/10 may be a test set. And training the user loss prediction model by using the training set, and verifying the user loss prediction model by using the verification set after each training round is finished. And when the accuracy reaches a preset threshold value, stopping training to obtain a trained user loss prediction model.

S404: the electronic equipment predicts the user training data through a gate control cycle unit and an attention mechanism module of the user loss prediction model to obtain a user loss training prediction result.

Specifically, as shown in fig. 3, the user training data is input into a user loss prediction model, a first training output hi is obtained through a gate control cycle unit of the user loss prediction model, a training attention probability distribution ei is obtained through the first training output by an attention mechanism module, the training attention probability distribution is processed to obtain a second training output yi, the second training output yi is input into a full connection layer to obtain a third training output, and the third training output is used for predicting the user data through an activation function to obtain a user loss training prediction result.

S406: and the electronic equipment updates parameters of the user loss prediction model according to the user loss training prediction result and the user loss historical data.

The training data comprises user training data and user loss historical data, the user training data can be used as input of the user loss prediction model, and parameters of the prediction model are updated through output of the user loss prediction model and the user loss historical data. Specifically, parameters of the user churn prediction model are updated according to the output of the user churn prediction model and a loss function of user churn historical data.

Based on the description of the above content, the present application provides a user churn prediction method, which obtains user data including time information through an electronic device, inputs the user data into a user churn prediction model, and predicts the user data through a gating cycle unit and an attention mechanism module of the user churn prediction model to obtain a user churn prediction result. Therefore, the user loss can be predicted through the user data comprising the time information, and the user data comprises the time information, so that the prediction can be performed based on the time characteristics, and the accuracy of the model is improved. In addition, the user loss prediction model introduces an attention mechanism, so that the importance degree of key features can be highlighted, and the accuracy of user loss prediction is further improved.

Corresponding to the above method embodiment, the present application further provides a user churn prediction apparatus, as shown in fig. 5, where the apparatus 500 includes: a communication module 502 and a prediction module 504.

A communication module for acquiring user data including time information;

In some possible implementations, the apparatus further includes:

The application provides equipment for realizing a verification method of an operation instruction. The apparatus includes a processor and a memory. The processor and the memory are in communication with each other. The processor is configured to execute instructions stored in the memory to cause the device to perform a user churn prediction method.

The present application provides a computer-readable storage medium having instructions stored thereon that, when run on a device, cause the device to perform the user churn prediction method described above.

A computer program product comprising instructions which, when run on a device, cause the device to perform the user churn prediction method described above.

It should be noted that the above-described embodiments of the apparatus are merely schematic, where the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. In addition, in the drawings of the embodiments of the apparatus provided in the present application, the connection relationship between the modules indicates that there is a communication connection therebetween, and may be implemented as one or more communication buses or signal lines.

Through the above description of the embodiments, those skilled in the art will clearly understand that the present application can be implemented by software plus necessary general-purpose hardware, and certainly can also be implemented by special-purpose hardware including special-purpose integrated circuits, special-purpose CPUs, special-purpose memories, special-purpose components and the like. Generally, functions performed by computer programs can be easily implemented by corresponding hardware, and specific hardware structures for implementing the same functions may be various, such as analog circuits, digital circuits, or dedicated circuits. However, for the present application, the implementation of a software program is more preferable. Based on such understanding, the technical solutions of the present application may be substantially embodied in the form of a software product, which is stored in a readable storage medium, such as a floppy disk, a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk of a computer, and includes several instructions for enabling a computer device (which may be a personal computer, an exercise device, or a network device) to execute the method according to the embodiments of the present application.

In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product.

The computer program product includes one or more computer instructions. When loaded and executed on a computer, cause the processes or functions described in accordance with the embodiments of the application to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, from one website site, computer, training device, or data center to another website site, computer, training device, or data center via wired (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can store or a data storage device, such as a training device, a data center, etc., that incorporates one or more available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.

Claims

1. A method for predicting user churn, the method comprising:

acquiring user data including time information;

2. The method of claim 1, wherein the user churn prediction model is obtained by training:

3. The method of claim 1, wherein the user data comprising time information comprises: and presetting all transaction behaviors of the user and corresponding time in the time.

4. The method of claim 3, wherein the user data further comprises: basic information data of the user.

5. The method of claim 1, wherein inputting the user data into a user churn prediction model, predicting the user data by a gated loop unit and an attention mechanism module of the user churn prediction model, and obtaining a user churn prediction result comprises:

6. The method of claim 1, wherein the attention mechanism module is located in a hidden layer of the user churn prediction model.

7. The method of any one of claims 1 to 6, further comprising:

8. A user churn prediction apparatus, the apparatus comprising:

a communication module for acquiring user data including time information;

9. An apparatus, comprising a processor and a memory;

the processor is to execute instructions stored in the memory to cause the device to perform the method of any of claims 1 to 7.

10. A computer-readable storage medium comprising instructions that direct a device to perform the method of any of claims 1-7.