User behavior recognition method and related equipment
By performing time sharding and feature extraction of user behavior data, combining variational autoencoder and bidirectional long and short-term memory networks, user behavior changes are captured, and the problem of difficult to accurately identify real-time changing user behavior in the prior art is solved, and efficient user behavior identification and credit risk assessment are achieved.
Patent Information
- Application Number
- CN202411753392.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-02
- Publication Date
- 2025-04-22
AI Technical Summary
The existing technology is difficult to accurately identify user behaviors that change in real time, and it depends on the experience and domain knowledge of data scientists to carry out complex feature engineering, which is time-consuming and labor-intensive. The model generalization ability is weak when dealing with new user behaviors, making it difficult to achieve accurate recognition.
By obtaining user behavior data in continuous time, time fragmentation is performed, user representation is extracted, behavioral features are extracted, and these features are input into the bidirectional long and short-term memory network to capture user behavior changes and generate behavioral change data.
It realizes accurate identification of user behavior, fully considers the dynamic changes in behavior characteristics, and deeply characterizes the behavioral relationships of users at different times, which improves the accuracy and effectiveness of credit risk assessment.
Smart Images
Figure CN119863306A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure generally relates to the field of electronic technology, and in particular to a user behavior recognition method, device, server, computer-readable storage medium, and computer program product. Background Art
[0002] In order to more accurately assess the credit risk of users and improve the accuracy and effectiveness of risk control of financial institutions, user behavior data (such as repayment records, consumption habits, and social network activities) can provide important evidence. At present, user behavior recognition usually adopts methods such as modeling user behavior data based on statistical theory, machine learning, and deep learning. These methods are difficult to accurately identify real-time changing user behaviors, and often rely on the experience and domain knowledge of data scientists to manually design complex feature engineering, which is time-consuming and labor-intensive. In addition, the model has weak generalization ability when dealing with new user behaviors, making it difficult to achieve accurate recognition. Summary of the invention
[0003] In view of the above-mentioned defects or deficiencies in the prior art, it is desired to provide a user behavior identification method, device, server, computer-readable storage medium and computer program product that can capture dynamic behavior information and achieve accurate behavior identification.
[0004] In a first aspect, an embodiment of the present application provides a method for identifying user behavior, including:
[0005] Acquire user behavior data within a continuous time, and perform time slicing on the user behavior data;
[0006] Extract user representation from user behavior data in each time slice;
[0007] Calling a variational autoencoder to extract behavioral features of the user representation corresponding to each time slice to obtain behavioral features in each time slice;
[0008] The behavioral features of multiple time slices are input into a bidirectional long short-term memory network, and user behavior changes are captured according to the correlation between different time slices to generate behavior change data.
[0009] In one embodiment, obtaining user behavior data within a continuous time period includes:
[0010] The user's click behavior data within the target APP within a preset time period is obtained as the user data.
[0011] In one embodiment, extracting user representations from user behavior data in each time slice includes:
[0012] Extract basic and statistical features of user behavior from the user behavior data in each time slice;
[0013] The basic features and the statistical features are combined as the user representation.
[0014] In one embodiment, the basic features include at least one of the number of clicks, number of active days, and most recent active time.
[0015] In one embodiment, the statistical features include at least one of active time distribution, click frequency, and event type concentration.
[0016] In one embodiment, it also includes:
[0017] Credit risk assessment is performed based on the behavior change data.
[0018] In a second aspect, an embodiment of the present application provides a user behavior recognition device, including:
[0019] A slicing processing unit, used to obtain user behavior data within a continuous time and perform time slicing on the user behavior data;
[0020] A representation extraction unit, used to extract user representations from user behavior data in each time slice;
[0021] A behavior feature extraction unit, used for calling a variational autoencoder to extract behavior features from the user representation corresponding to each time slice, so as to obtain behavior features in each time slice;
[0022] The behavior change capture unit is used to input the behavior features of multiple time slices into a bidirectional long short-term memory network, capture user behavior changes according to the correlation between different time slices, and generate behavior change data.
[0023] In a third aspect, an embodiment of the present application provides a server, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method described in the embodiment of the present application when executing the program.
[0024] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described in the embodiment of the present application.
[0025] In a fifth aspect, an embodiment of the present application provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of the method described in the embodiment of the present application.
[0026] Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description, or will be learned through practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:
[0028] Figure 1 A schematic diagram of a process flow of a user behavior identification method provided in an embodiment of the present application is shown;
[0029] Figure 2 A schematic diagram of a user behavior identification process provided by an embodiment of the present application is shown;
[0030] Figure 3 An exemplary structural block diagram of a user behavior identification device provided in an embodiment of the present application is shown;
[0031] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing a server of an embodiment of the present application is shown. DETAILED DESCRIPTION
[0032] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It is also necessary to explain that, for ease of description, only the parts related to the invention are shown in the accompanying drawings.
[0033] It should be noted that, in the absence of conflict, the embodiments in the present application and the features in the embodiments can be combined with each other. The present application will be described in detail below with reference to the drawings and in combination with the embodiments. Although the embodiments of the present application provide the method operation instruction steps shown in the following embodiments or drawings, more or fewer operation instruction steps may be included in the method based on routine or no creative labor. In the steps where there is no necessary causal relationship logically, the execution order of these steps is not limited to the execution order provided by the embodiments of the present application. The method may be executed in the order of the methods shown in the embodiments or drawings or in parallel during the actual processing process or when the device is executed.
[0034] Please refer to Figure 1 , Figure 1 FIG. 1 is a flow chart of a method for identifying user behavior provided by an embodiment of the present application. Figure 1 As shown, the method includes:
[0035] S101, obtaining user behavior data within a continuous time period, and performing time slicing on the user behavior data;
[0036] The collected user behavior data within a continuous time is time-sliced to obtain user behavior data of several slices. There is no limit on the time interval of the specific slices, which can be set accordingly according to the actual analysis needs.
[0037] By analyzing user behavior data over continuous time, user behavior is deeply characterized based on the correlation between user behaviors at different times to achieve accurate behavior identification.
[0038] The types of user behavior data obtained, such as user visits to pages, clicks, searches, communication with customer service, posting, commenting, liking, etc. in communities or forums, the number of push notifications received, the number of push notification clicks, etc., are not limited here. Optionally, the user's click behavior data in the target APP within a preset time period can be obtained as user data. Among user behavior data, click behavior is the most direct way of user interaction, and the acquisition method is relatively simple. For example, the user's click behavior is collected through embedding points in the target APP to generate click behavior data. Click behavior reflects the user's interest and preference for different content and functions. By analyzing the user's click behavior, its potential usage habits, interest changes, and even abnormal behavior can be discovered, which are helpful for user risk identification. Of course, in addition to user click behavior data, other data types can also be added for other analysis needs, which will not be repeated here.
[0039] It should be noted that the acquisition and use of the above information requires the user's consent and can only be used after the user's authorization.
[0040] S102, extracting user representation from user behavior data in each time slice;
[0041] Analyze the user behavior data in each shard, divide the continuous user behavior into different time periods by taking different window sizes, perform feature statistics on the user performance in each time period, extract user representation, and form the initial embedding of user behavior. User representation refers to data used to describe and represent user characteristics, preferences, behavior patterns, and other information in a computer system. Extract user representation to better understand and predict user behavior. The specific information type for extracting user representation is not limited in this embodiment.
[0042] Optionally, basic features and statistical features of the user behavior may be extracted from the user behavior data in each time slice; the basic features and statistical features may be combined as user representation.
[0043] Among them, basic features refer to the original information that can be directly obtained from user data. This information is usually static or semi-static and reflects the inherent attributes or basic information of the user. Basic features include but are not limited to the following categories: basic personal information (age, gender, education background, occupation, income level), account information (registration time, registration method), device information (device type, operating system, device model), etc. Optionally, basic features can specifically include: at least one of the number of clicks, number of active days, and the last active time; statistical features refer to the features obtained by statistical analysis of user behavior data. These features reflect the user's behavior patterns and trends, and are usually dynamic and sequential. Statistical features include but are not limited to the following categories: behavior frequency (login frequency, transaction frequency, page browsing frequency), behavior intensity (average duration of each login, average number of pages browsed each time), behavior pattern (common transaction time period, commonly used payee), etc. Optionally, statistical features can specifically include: at least one of active time distribution, click frequency, and event type concentration. Basic features provide static information of users, while statistical features reflect the dynamic behavior patterns of users. Combining basic features and statistical features can more comprehensively assess the credit risk of users.
[0044] S103, calling a variational autoencoder to extract behavioral features from the user representation corresponding to each time slice, to obtain behavioral features within each time slice;
[0045] The user behavior features within a fixed time period are extracted through the variational autoencoder. After encoding and decoding by the variational autoencoder, the key behavior features of the user behavior in each time slice are obtained.
[0046] The behavior feature extraction through variational autoencoder is mainly divided into two parts: encoder and decoder. Figure 2 The figure shows a schematic diagram of the user behavior recognition process, where the encoder consists of a two-layer MLP network that maps simple statistical features to the latent space. The output of the encoder is the mean and variance of the latent variable, which is used to describe the distribution in the latent space, so that the latent variable can be sampled from the distribution. The decoder also consists of a two-layer MLP network, which reconstructs the data by processing the output of the encoder and learns the mapping from the latent space to the original data space, so that the reconstructed data is as close as possible to the original input data.
[0047] By reconstructing user behavior through a variational autoencoder, we can capture the key information in the initial embedding while removing noise and redundancy. The features processed by the variational autoencoder are more diverse based on the noise removal of the initial features. It should be noted that if there are fewer behavioral features, this method can be used for data enhancement.
[0048] S104: input the behavior features of multiple time slices into a bidirectional long short-term memory network, capture the user behavior changes according to the correlation between different time slices, and generate behavior change data.
[0049] The bidirectional long short-term memory network takes multiple consecutive window periods, that is, the behavioral feature sequences in multiple slices as input, and selectively memorizes the information in the sequence through the learning of the gating mechanism, thereby capturing the changing pattern of user behavior. Among them, the forward network is responsible for reading the sequence from front to back and learning the positive dependency of the sequence, and the backward network reads the behavior sequence from back to front and captures the reverse dependency of the sequence, thereby realizing the extraction of contextual association information of the behavior sequence.
[0050] The bidirectional long short-term memory network learns the user's behavior sequence in multiple time periods. It can learn the dependencies within the sequence while capturing the user's behavior and interest and willingness to participate in financial activities. Combined with the bidirectional long short-term memory network, it learns the correlation between user behavior in different time periods, thereby deeply representing user behavior.
[0051] In one embodiment, after generating accurate behavior change data, credit risk assessment can be further performed based on the behavior change data to achieve accurate financial risk prediction. Specifically, abnormal changes (such as large amounts of borrowing in a short period of time, etc.) can be identified based on the behavior change data, and credit warning rules corresponding to various abnormal changes can be set to achieve credit risk assessment. It should be noted that the specific implementation process of specific credit risk assessment in this embodiment is not limited, and can be set accordingly according to the actual needs of user behavior monitoring. Here, only the above process is used as an example for introduction.
[0052] It should be noted that although the operations of the method of the present invention are described in a particular order in the drawings, this does not require or imply that the operations must be performed in this particular order or that all illustrated operations must be performed to achieve desired results.
[0053] Based on the above introduction, the method provided in this embodiment performs time slicing on the acquired user behavior data in continuous time, calculates the user representation in each slicing, and calls the variational autoencoder to extract the behavior features of the user representation, which can extract key information while enhancing sample diversity, and then capture the changes in user behavior according to the correlation between different time slices through the bidirectional long short-term memory network, thereby achieving a deep representation of user behavior. Compared with the traditional model, this method fully considers the dynamic changes of behavioral characteristics, deeply represents user behavior according to the behavioral correlation between users at different times, and can achieve accurate behavior recognition.
[0054] The above embodiments do not limit the training process of the variational autoencoder and the bidirectional long short-term memory network. To deepen the understanding, this embodiment introduces a training method of a model, and the details can be referred to the following introduction.
[0055] The training objective of the model consists of two stages, the key information extraction stage and the temporal learning stage. In the key information extraction stage, the training objective is expressed as Loss VAE, which consists of reconstruction loss Reconstruction Loss and KL divergence; the former mainly measures the difference between the reconstructed sample and the initial embedding of the user behavior. Ideally, the decoder is expected to generate samples that are as similar as possible to the original data. KL divergence measures the difference between the distribution in the latent space and the standard normal distribution, encouraging points in the latent space to be distributed near a standard normal distribution, making the sampling process more efficient. This helps the model learn a smooth latent space representation and improve the diversity of generated samples. Loss VAE is expressed as follows:
[0056] L VAE =L reconstruction +β·L KL
[0057]
[0058] Among them, q(z|x) is the approximate distribution in the latent space, p(z) is the prior distribution, and usually the prior distribution is assumed to be the standard normal distribution N(0,I). i and σ i are the mean and variance of the latent variable z, d is the dimension of the latent space, and β is a hyperparameter used to balance the reconstruction loss and the KL divergence term.
[0059] In the time series learning stage, the training target is the cross entropy loss function, where y is the true label and y^ is the predicted probability.
[0060]
[0061] Taking the two stages into consideration, the overall training goal of the model is composed of the goals of the two stages, which can be expressed as
[0062] L joint =L VAE +λ·L cross-entropy
[0063] Among them, λ is a hyperparameter used to balance the VAE loss and the cross entropy loss.
[0064] Further references Figure 3 , which shows an exemplary structural block diagram of a user behavior recognition device according to an embodiment of the present application, which mainly includes:
[0065] A slicing processing unit 101 is used to obtain user behavior data within a continuous time and perform time slicing on the user behavior data;
[0066] A representation extraction unit 102, configured to extract user representations from user behavior data in each time slice;
[0067] The behavior feature extraction unit 103 is used to call the variational autoencoder to extract the behavior features of the user representation corresponding to each time slice, so as to obtain the behavior features in each time slice;
[0068] The behavior change capturing unit 104 is used to input the behavior features of multiple time slices into the bidirectional long short-term memory network, capture the user behavior changes according to the correlation between different time slices, and generate behavior change data.
[0069] It should be understood that the units described in the above device are similar to those in the reference Figure 1 The steps in the method described above correspond to each other. Therefore, the operations and features described above for the method are also applicable to the device and the units contained therein, and will not be repeated here. The device can be pre-implemented in the browser or other security application of the server, or loaded into the browser or its security application of the server by downloading or the like. The corresponding units in the device can cooperate with the units in the server to implement the solution of the embodiment of the present application.
[0070] For the several units mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiments of the present disclosure, the features and functions of two or more units described above can be embodied in one unit. On the contrary, the features and functions of one unit described above can be further divided into multiple units to be embodied.
[0071] It should be noted that for details not disclosed in the user behavior identification device in the embodiment of the present application, please refer to the details disclosed in the above embodiments of the present application, which will not be repeated here.
[0072] Reference below Figure 4 , Figure 4 A schematic diagram of the structure of a computer system suitable for implementing a server of an embodiment of the present application is shown.
[0073] like Figure 4As shown, the computer system includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 402 or the program loaded from the storage part 408 into the random access memory (RAM) 403. In the RAM 403, various programs and data required for the operation instructions of the system are also stored. The CPU 401, the ROM 402, and the RAM 403 are connected to each other through a bus 404. An input / output (I / O) interface 405 is also connected to the bus 404.
[0074] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, a mouse, etc.; an output section 407 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN card, a modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. A removable medium 411, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 410 as needed, so that a computer program read therefrom is installed into the storage section 408 as needed.
[0075] In particular, according to an embodiment of the present application, the above reference flow chart Figure 1 The described process can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program includes a program code for executing the method shown in the flow chart. In such an embodiment, the computer program includes a program code for executing the method shown in the flow chart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 409, and / or installed from the removable medium 411. When the computer program is executed by the central processing unit (CPU) 401, the above-mentioned functions defined in the system of the present application are executed.
[0076] It should be noted that the computer-readable medium shown in the present application can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium such as a computer-readable storage medium that can send, propagate or transmit a program for use by or in conjunction with an instruction execution system, device or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.
[0077] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, functions and operating instructions of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the aforementioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some implementations as replacements, the functions marked in the box can also occur in a sequence different from that marked in the accompanying drawings. For example, the boxes represented by two connections can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operating instruction, or can be implemented with a combination of dedicated hardware and computer instructions.
[0078] The units or modules involved in the embodiments described in the present application may be implemented by software or hardware. The units or modules described may also be arranged in a processor. The names of these units or modules do not, in some cases, constitute limitations on the units or modules themselves.
[0079] As another aspect, the present application further provides a computer-readable storage medium, which may be included in the server described in the above embodiment, or may exist independently without being assembled into the server. The above computer-readable storage medium stores one or more programs, and when the above programs are used by one or more processors to execute the user behavior identification method described in the present application.
[0080] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of disclosure involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the aforementioned disclosed concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.
Claims
1. A user behavior identification method, characterized in that: include: Acquire user behavior data within a continuous time, and perform time slicing on the user behavior data; Extract user representation from user behavior data in each time slice; Calling a variational autoencoder to extract behavioral features of the user representation corresponding to each time slice to obtain behavioral features in each time slice; The behavioral features of multiple time slices are input into a bidirectional long short-term memory network, and user behavior changes are captured according to the correlation between different time slices to generate behavior change data.
2. The method according to claim 1, characterized in that The obtaining of user behavior data over a continuous period of time includes: The user's click behavior data in the target APP within a preset time period is obtained as the user behavior data.
3. The method according to claim 1, characterized in that The extracting of user representations from user behavior data in each time slice includes: Extract basic and statistical features of user behavior from the user behavior data in each time slice; The basic features and the statistical features are combined as the user representation.
4. The method according to claim 3, characterized in that The basic features include: at least one of the number of clicks, active days, and the most recent active time.
5. The method according to claim 3, characterized in that The statistical features include: at least one of active time distribution, click frequency, and event type concentration.
6. The method according to claim 1, characterized in that Also includes: Credit risk assessment is performed based on the behavior change data.
7. A user behavior recognition device, characterized in that: include: A slicing processing unit, used to obtain user behavior data within a continuous time and perform time slicing on the user behavior data; A representation extraction unit, used to extract user representations from user behavior data in each time slice; A behavior feature extraction unit, used for calling a variational autoencoder to extract behavior features from the user representation corresponding to each time slice, so as to obtain behavior features in each time slice; The behavior change capture unit is used to input the behavior features of multiple time slices into a bidirectional long short-term memory network, capture user behavior changes according to the correlation between different time slices, and generate behavior change data.
8. A server comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the method according to any one of claims 1 to 6 are implemented.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.