Data recommendation method and device and electronic equipment

By extracting different user behavior characteristics whose correlation with historical user behavior characteristics is less than the set correlation, and training user interest prediction models, the problem that traditional recommendation systems are difficult to capture changes in user interest is solved, and the accuracy and real-timeness of recommendations are improved.

CN119089038APending Publication Date: 2024-12-06BEIYIN FINANCIAL TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411095255.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Traditional algorithmic recommendation systems are difficult to capture changes in user interests, resulting in low recommendation accuracy and poor real-time performance, especially when there is a big difference between user's actual preferences and the recommended content of registration information.

Method used

By obtaining the current user behavior data, differential user behavior characteristics with less correlation with historical user behavior characteristics than the set correlation degree are extracted. Based on these characteristics, the user interest prediction model is trained, and the recommended results are consistent with the user's latest interest preferences are output, and according to the user's user's current interest preferences, Feedback to adjust the recommendation strategy.

Benefits of technology

It realizes real-time adjustment of recommendation strategies based on changes in user behavior, improves the accuracy and real-timeness of recommendations, and ensures that the recommended content meets users' latest needs and interests.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119089038A_ABST
    Figure CN119089038A_ABST
Patent Text Reader

Abstract

The invention provides a data recommendation method and device and electronic equipment, and relates to the technical field of data processing. The method comprises the following steps: firstly, acquiring current user behavior data, and selecting a different user behavior feature from a feature set corresponding to the current user behavior data, wherein the association degree between the different user behavior feature and a historical user behavior feature is smaller than a set association degree; then training a preset user interest prediction model based on the differential user behavior characteristics and the historical user behavior characteristics; and finally, outputting a recommendation result corresponding to the interest preference of the user through the user interest prediction model. In this way, it can be ensured that the recommended content always meets the newest requirements and interests of the user, and the accuracy of content recommendation is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a data recommendation method, device and electronic device. Background Art

[0002] With the rapid development of Internet technology, computer networks have been widely used in various fields of people's lives, such as online shopping, watching videos and browsing media information.

[0003] With the growth of Internet information, users hope to dig out information content that meets their interests and preferences from massive information resources. Therefore, it is very important and urgent to accurately explore user preferences and make efficient and accurate content recommendations.

[0004] Traditional algorithm recommendation systems, such as collaborative filtering systems, usually generate recommendations based on the interaction information between users and recommended content, and can recommend relevant content to users based on user preference tags. However, for user preference tags that have never appeared in the algorithm recommendation system or have not been trained by the algorithm recommendation system, the algorithm recommendation system cannot predict the content that users prefer.

[0005] Specifically, when an algorithm recommendation system recommends content that users like, it usually determines the target information that users are interested in based on the relevant registration information (such as interest tags) filled in by the users, and recommends the target information to the users. However, when there is a large difference between the content that users actually like and the target information recommended by the algorithm recommendation system based on the registration information, it is difficult for the algorithm recommendation system to capture the changes in users' interests, resulting in low recommendation accuracy and poor real-time performance. Summary of the invention

[0006] The present invention provides a data recommendation method, device and electronic device for recommending content that matches the user's interest changes in real time, thereby improving the accuracy of the recommendation. The specific technical solution is as follows:

[0007] In a first aspect, the present application provides a data recommendation method, comprising:

[0008] Acquire current user behavior data, and select, from a feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with a historical user behavior feature is less than a set correlation;

[0009] Based on the difference user behavior characteristics and the historical user behavior characteristics, a preset user interest prediction model is trained;

[0010] The user interest prediction model is used to output recommendation results corresponding to the user's interest preferences.

[0011] Based on the above method, the recommendation strategy can be adjusted in real time according to changes in user behavior to ensure that the recommended content meets the user's latest needs and interests, thereby improving the accuracy of the recommendation.

[0012] In a possible implementation, obtaining current user behavior data includes:

[0013] Obtain initial user behavior data;

[0014] Performing data cleaning and denoising on the initial user behavior data to obtain target user behavior data;

[0015] The target user behavior data is compressed and desensitized to obtain the current user behavior data.

[0016] Based on the above method, the initial user behavior data can be obtained, and by performing data cleaning, denoising, compression, desensitization and encryption on the initial user behavior data, a rich and effective amount of current user behavior data can be obtained to prepare data for subsequent model training.

[0017] In a possible implementation, selecting, from the feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with the historical user behavior feature is less than a set correlation includes:

[0018] Based on a deep learning network, extract a feature set corresponding to the current user behavior data;

[0019] Calculating the correlation between each feature in the feature set and the historical user behavior feature;

[0020] Based on the correlation between each feature and the historical user behavior feature, each feature is sorted and the top n features are selected;

[0021] The n features are used as the difference user behavior features, where n is an integer greater than 0.

[0022] Based on the above method, it is possible to select different user behavior features that have a large change from historical user behavior features from the feature set of current user behavior data.

[0023] In a possible implementation, taking the n features as the difference user behavior features includes:

[0024] Combining the n features according to their relevance to obtain m new features;

[0025] The m newly added features and the n features are used as the difference user behavior features, where m is an integer greater than 0.

[0026] Based on the above method, n features can be combined according to relevance to generate m new features, thereby increasing the sample size of the user interest prediction model and improving the generalization ability of the user interest prediction model.

[0027] In a possible implementation, after outputting a recommendation result corresponding to the user's interest preference through the user interest prediction model, the method further includes:

[0028] Displaying the recommendation results on the homepage according to the set recommendation priority;

[0029] Based on the user's homepage operations on the recommended homepage, the layout, theme or style of the recommended homepage is adjusted, and according to the feedback data returned by the user for the recommendation results, the recommendation strategy of the recommended homepage is adjusted in combination with reinforcement learning or online learning algorithms.

[0030] Based on the above method, it can ensure that the recommended content always meets the user's latest needs and interests, meet the diverse needs of different users for recommended homepages, and enrich the user experience.

[0031] In a second aspect, the present application provides a data recommendation device, comprising:

[0032] A data acquisition module, used to acquire current user behavior data, and select, from a feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with a historical user behavior feature is less than a set correlation;

[0033] A model training module, used for training a preset user interest prediction model based on the differential user behavior characteristics and the historical user behavior characteristics;

[0034] The result generation module is used to output the recommendation results corresponding to the user's interest preferences through the user interest prediction model.

[0035] In a possible implementation, the data acquisition module is specifically used to:

[0036] Based on a deep learning network, extract a feature set corresponding to the current user behavior data;

[0037] Calculating the correlation between each feature in the feature set and the historical user behavior feature;

[0038] Based on the correlation between each feature and the historical user behavior feature, each feature is sorted and the top n features are selected;

[0039] The n features are used as the difference user behavior features, where n is an integer greater than 0.

[0040] In a possible implementation, the device further includes a recommendation display module, and the recommendation display module is specifically configured to:

[0041] Displaying the recommendation results on the homepage according to the set recommendation priority;

[0042] Based on the user's homepage operations on the recommended homepage, the layout, theme or style of the recommended homepage is adjusted, and according to the feedback data returned by the user for the recommendation results, the recommendation strategy of the recommended homepage is adjusted in combination with reinforcement learning or online learning algorithms.

[0043] In a third aspect, the present application provides an electronic device, including:

[0044] Memory, used to store computer programs;

[0045] The processor is used to implement the steps of the above-mentioned data recommendation method when executing the computer program stored in the memory.

[0046] In a fourth aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned data recommendation method are implemented.

[0047] For each aspect from the second to the fourth aspect and the technical effects that may be achieved by each aspect, please refer to the above description of the technical effects that can be achieved by the first aspect or various possible schemes in the first aspect, and no further details will be given here. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 A flowchart of a data recommendation method provided in an embodiment of the present application;

[0049] Figure 2 A schematic diagram of the data recommendation system architecture provided in an embodiment of the present application;

[0050] Figure 3 A schematic diagram of the structure of a data recommendation device provided in an embodiment of the present application;

[0051] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The specific operating methods in the method embodiments can also be applied to device embodiments or system embodiments. It should be noted that in the description of the present application, "multiple" is understood as "at least two". "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A is connected to B, which can represent: A is directly connected to B and A is connected to B through C. In addition, in the description of the present application, words such as "first" and "second" are only used to distinguish the purpose of description, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0053] The embodiments of the present application are described in detail below in conjunction with the accompanying drawings.

[0054] With the rapid development of Internet technology, computer networks have been widely used in various fields of people's lives, such as online shopping, watching videos and browsing media information.

[0055] With the growth of Internet information, users hope to dig out information content that meets their interests and preferences from massive information resources. Therefore, it is very important and urgent to accurately explore user preferences and make efficient and accurate content recommendations.

[0056] Traditional algorithm recommendation systems, such as collaborative filtering systems, usually generate recommendations based on the interaction information between users and recommended content, and can recommend relevant content to users based on user preference tags. However, for user preference tags that have never appeared in the algorithm recommendation system or have not been trained by the algorithm recommendation system, the algorithm recommendation system cannot predict the content that users prefer.

[0057] Specifically, when an algorithm recommendation system recommends content that users like, it usually determines the target information that users are interested in based on the relevant registration information (such as interest tags) filled in by the users, and recommends the target information to the users. However, when there is a large difference between the content that users actually like and the target information recommended by the algorithm recommendation system based on the registration information, it is difficult for the algorithm recommendation system to capture the changes in users' interests, resulting in low recommendation accuracy and poor real-time performance.

[0058] In view of this, in order to recommend content that is in line with changes in interests to users in real time and improve the accuracy of recommendations, the present application provides a data recommendation method, which specifically includes: first, obtaining current user behavior data, and selecting, from a feature set corresponding to the current user behavior data, differential user behavior features whose correlation with historical user behavior features is less than a set correlation; then, based on the differential user behavior features and the historical user behavior features, training a preset user interest prediction model; finally, through the user interest prediction model, outputting a recommendation result corresponding to the user's interest preference.

[0059] Through the method provided in this application, the recommendation strategy can be adjusted in real time according to changes in user behavior, ensuring that the recommended content meets the user's latest needs and interests, thereby improving the accuracy of the recommendation.

[0060] Reference Figure 1 As shown, it is a flow chart of a data recommendation method provided in an embodiment of the present application, the method comprising:

[0061] S1, obtaining current user behavior data, and selecting, from a feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with the historical user behavior feature is less than a set correlation.

[0062] First of all, the method provided in this application can be applied to Figure 2 In the system architecture shown, the system architecture includes: a terminal device, and the method provided in the present application can be executed in the terminal device.

[0063] The terminal device can be a computer, smart phone, tablet or other intelligent terminal. The terminal device can record the user's behavior operations and recommend content that meets the user's latest needs and interests based on the user's behavior operations.

[0064] In the embodiment of the present application, the terminal device can obtain the current user behavior data through various channels. Specifically, the controller can obtain the user's registration information, browsing history, search history, purchase history and other initial user behavior data through browser plug-ins, APP tracking, social media interfaces, etc.

[0065] In order to remove redundant information and erroneous information in the initial user behavior data, the controller can perform data cleaning and denoising on the initial user behavior data to obtain target user behavior data. In specific implementation, the terminal device can use a distributed big data processing platform to perform data cleaning and denoising to obtain target user behavior data. This application does not impose any specific restrictions on this.

[0066] After obtaining the target user behavior data, the terminal device can compress and desensitize the target user behavior data to obtain the current user behavior data. In addition, in order to protect the user privacy in the target user behavior data and ensure the security and confidentiality of the user privacy, after obtaining the target user behavior data, the terminal device can also encrypt the target user behavior data by using differential hiding technology or end-to-end encryption.

[0067] Through the above method, the initial user behavior data can be obtained, and by performing data cleaning, denoising, compression, desensitization and encryption on the initial user behavior data, a rich and effective amount of current user behavior data can be obtained to prepare data for subsequent model training.

[0068] In an embodiment of the present application, after obtaining the current user behavior data, the terminal device can select, from the feature set corresponding to the current behavior data, a difference user behavior feature whose correlation with the historical user behavior feature is less than a set correlation.

[0069] Specifically, after obtaining the current user behavior data, the terminal device can use a deep learning network, such as a convolutional neural network model (CNN), a recurrent neural network model, etc., to extract a feature set corresponding to the current user behavior data. The feature set may include the user's interest tags, behavior patterns, emotional tendencies, etc., where the interest tags may be the types of TV programs that the user is interested in, such as sports, movies, TV series, news information, etc. This application does not impose any specific restrictions on this.

[0070] The behavior pattern can be progressive, reciprocating, or random. Progressive refers to the behavior when the user's goal is clear. For example, when a user buys a mobile phone, the user's behavior path at this time is to open the shopping app - search for the phone name - browse the search results - select the product - browse the product details - add to the shopping cart - pay. This behavior path is linear, that is, progressive. The reciprocating and random types will not be described in detail here.

[0071] Sentimental tendency refers to the user's emotional attitude towards a specific subject or event, for example, the emotions, attitudes, and opinions expressed when making comments on social networks, including positive, negative, or neutral.

[0072] In an embodiment of the present application, when the terminal device obtains the feature set corresponding to the current user behavior data, it can combine the knowledge graph to build a multi-dimensional user portrait for the user, where the user portrait can include the user's basic attributes, interests and hobbies, consumption habits, social relationships, etc. The terminal device can push the latest user-preferred recommended content to the user based on the user portrait.

[0073] When the terminal device obtains the feature set corresponding to the current user behavior data, it can calculate the correlation between each feature in the feature set and the historical user behavior feature. The historical user behavior feature represents the historical behavior feature extracted from the user's historical user behavior data. The method for obtaining the historical behavior feature can refer to the method for obtaining the feature set corresponding to the current user behavior data mentioned above, which will not be repeated here.

[0074] The terminal device can determine the degree of difference between the user's current interest preferences and historical interest preferences based on the correlation between each feature in the feature set and the historical behavior feature; if the correlation is low, it means that the user's current interest preferences and historical interest preferences have changed significantly. In order to recommend content that meets the user's current interest preferences, it is necessary to train and update the interest recommendation model based on the changed user behavior characteristics.

[0075] In the embodiment of the present application, the terminal device may first sort each feature in the feature set based on the degree of association between each feature in the feature set and the historical user behavior feature, and then select the top n features. In specific implementation, the features may be sorted in reverse order according to the degree of association, and then the top n features may be selected, and finally these n features are used as the differential user behavior features. The present application does not impose any specific restrictions on the arrangement method.

[0076] It should be noted that n is an integer greater than 0.

[0077] Through the above method, it is possible to select different user behavior features that have a large change from historical user behavior features from the feature set of current user behavior data.

[0078] In a possible implementation, in order to increase the training set of the interest recommendation model and improve the generalization ability of the interest recommendation model, after obtaining n features, the terminal device can also combine the n features according to the correlation to obtain m new features. For example, two features with high correlation among the n features can be combined in pairs to obtain m new features, where m is an integer greater than 0.

[0079] After obtaining the m newly added features, the controller may use the m newly added features and the n features as differential user behavior features, and the differential user behavior features may be used to characterize changes in user interest preferences.

[0080] Through the above method, it is possible to mine user behavior patterns and behavior characteristics from massive data, thereby more accurately describing user characteristics and providing a solid foundation for predicting recommended content based on the user's latest interest preferences.

[0081] S2, based on the differential user behavior characteristics and historical user behavior characteristics, the preset user interest prediction model is trained.

[0082] In an embodiment of the present application, after obtaining the differential user behavior characteristics, the terminal device can train a preset user interest prediction model based on the differential user behavior characteristics and the historical user behavior characteristics. In the specific implementation, a machine learning algorithm, such as collaborative filtering, deep learning, etc., can be used to train the user interest prediction model.

[0083] In addition, in the embodiment of the present application, in order to take into account the real-time nature of model training, the model parameters are optimized. After obtaining the current user behavior data, the terminal device can use real-time data processing technology to process the current user behavior data through pre-set scheduling prevention. For example, the current user behavior data is automatically detected, associated, correlated, estimated, and combined to obtain differential user behavior characteristics; then, the time series analysis technology is used to perform trend analysis and residual analysis on the differential user behavior characteristics, wherein the residual analysis of the differential user behavior characteristics can filter out abnormal user behavior characteristics or user behavior characteristics that are not related to user interest preferences in the differential user behavior characteristics; the residual analysis can infer the user's interest preferences.

[0084] Through the above method, after obtaining the differential user behavior characteristics of the current user behavior data, the differential user behavior characteristics and historical user behavior characteristics can be used to perform real-time training on the user interest prediction model, so that the output of the user interest prediction model meets the recommended content of the user's latest interest preferences.

[0085] S3, through the user interest prediction model, output the recommendation results corresponding to the user's interest preferences.

[0086] In the implementation of this application, after training the user interest prediction model, the terminal device can generate a trained user interest prediction model, and output recommendation results corresponding to the user's interest preferences through the trained user interest prediction model. The recommendation results can meet the user's latest interest preference corresponding to the recommended content needs, and improve the accuracy and real-time performance of the prediction of recommended content.

[0087] In the embodiment of the present application, after the user interest prediction model outputs the recommendation results, the terminal device can display the recommendation results on the homepage according to the set recommendation priority. Specifically, the terminal device can use the attention mechanism and combine the time series analysis technology to insert the first priority recommendation results in the key position (eye-catching position) of the homepage, and insert the second priority recommendation results in the marginal area (such as the footer), wherein the first priority is greater than the second priority.

[0088] In addition, in order to facilitate users to adjust the layout, color theme, font style, etc. of the recommended homepage according to their personal preferences, users can perform corresponding homepage operations on the recommended homepage, such as theme selection operations and font style operations in the homepage settings, so that the recommended homepage meets the user's interests and preferences. In specific implementation, technologies such as G2 and X6 can be used to provide users with rich interface templates and customization options, support user drag-and-drop operations, and make the recommended homepage meet the user's personalized needs.

[0089] In an embodiment of the present application, in order to optimize the model, the terminal device can also adjust the recommendation strategy of the recommended homepage based on the feedback data returned by the user regarding the recommendation results, combined with reinforcement learning or online learning algorithms; specifically, during use, the user can provide feedback on the recommendation results through operations such as likes, shares, comments, and hiding. After collecting these feedback data, the terminal device can use reinforcement learning or online learning algorithms to adjust and optimize the recommendation strategy to achieve continuous optimization of the customization effect.

[0090] In summary, the method provided in this application utilizes advanced algorithms such as deep learning to mine user behavior patterns and behavioral characteristics from massive data, thereby more accurately describing user characteristics and providing a solid foundation for personalized recommendations; by collecting and analyzing multi-dimensional information such as user browsing history, click behavior, purchase history, social media data, etc., it is possible to construct a detailed and accurate user portrait and predict the user's subsequent recommended content; by extracting differential user behavior data in real time for training, and adjusting the recommendation strategy of the recommendation homepage based on the feedback data returned by the user for the recommendation results, it is possible to ensure that the recommended content always meets the user's latest needs and interests, thereby improving the accuracy of content recommendations.

[0091] Based on the method provided in the above embodiment, the embodiment of the present application also provides a data recommendation device, such as Figure 3 The figure is a schematic diagram of the structure of a data recommendation device in an embodiment of the present application, the device comprising:

[0092] The data acquisition module 301 is used to acquire current user behavior data, and select, from a feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with the historical user behavior feature is less than a set correlation;

[0093] A model training module 302, used to train a preset user interest prediction model based on the difference user behavior characteristics and the historical user behavior characteristics;

[0094] The result generation module 303 is used to output a recommendation result corresponding to the user's interest preference through the user interest prediction model.

[0095] In a possible implementation, the data acquisition module 301 is specifically used for:

[0096] Based on a deep learning network, extract a feature set corresponding to the current user behavior data;

[0097] Calculating the correlation between each feature in the feature set and the historical user behavior feature;

[0098] Based on the correlation between each feature and the historical user behavior feature, each feature is sorted and the top n features are selected;

[0099] The n features are used as the difference user behavior features, where n is an integer greater than 0.

[0100] In a possible implementation, the device further includes a recommendation display module 304, and the recommendation display module is specifically configured to:

[0101] Displaying the recommendation results on the homepage according to the set recommendation priority;

[0102] Based on the user's homepage operations on the recommended homepage, the layout, theme or style of the recommended homepage is adjusted, and according to the feedback data returned by the user for the recommendation results, the recommendation strategy of the recommended homepage is adjusted in combination with reinforcement learning or online learning algorithms.

[0103] Based on the same inventive concept, an electronic device is also provided in an embodiment of the present application, and the electronic device can realize the functions of the aforementioned data recommendation device, referring to Figure 4 , the electronic device comprises:

[0104] At least one processor 401, and a memory 402 connected to the at least one processor 401. The specific connection medium between the processor 401 and the memory 402 is not limited in the embodiment of the present application. Figure 4 In the example, the processor 401 and the memory 402 are connected via the bus 400. The bus 400 is Figure 4 The connection between other components is shown by bold lines, and is not intended to be limiting. The bus 400 can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 4 Only one thick line is used in the figure, but it does not mean that there is only one bus or one type of bus. Alternatively, the processor 401 can also be called a controller, and there is no limitation on the name.

[0105] In the embodiment of the present application, the memory 402 stores instructions that can be executed by at least one processor 401. The at least one processor 401 can execute the data recommendation method discussed above by executing the instructions stored in the memory 402. The processor 401 can implement Figure 3 The functions of each module in the device shown.

[0106] Among them, the processor 401 is the control center of the device, and can use various interfaces and lines to connect the various parts of the entire control device. By running or executing instructions stored in the memory 402 and calling the data stored in the memory 402, the various functions of the device and process data, the device can be monitored as a whole.

[0107] In one possible design, the processor 401 may include one or more processing units, and the processor 401 may integrate an application processor and a modem processor, wherein the application processor mainly processes an operating system, a user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the modem processor may not be integrated into the processor 401. In some embodiments, the processor 401 and the memory 402 may be implemented on the same chip, and in some embodiments, they may also be implemented separately on separate chips.

[0108] Processor 401 can be a general-purpose processor, such as a central processing unit (CPU), a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component, and can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. A general-purpose processor can be a microprocessor or any conventional processor, etc. The steps of the data recommendation method disclosed in the embodiments of the present application can be directly embodied as a hardware processor, or can be executed by a combination of hardware and software modules in the processor.

[0109] The memory 402 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules. The memory 402 may include at least one type of storage medium, such as a flash memory, a hard disk, a multimedia card, a card-type memory, a random access memory (Random Access Memory, RAM), a static random access memory (Static Random Access Memory, SRAM), a programmable read-only memory (Programmable Read Only Memory, PROM), a read-only memory (Read Only Memory, ROM), an electrically erasable programmable read-only memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), a magnetic memory, a disk, an optical disk, etc. The memory 402 is any other medium that can be used to carry or store a desired program code in the form of an instruction or data structure and can be accessed by a computer, but is not limited thereto. The memory 402 in the embodiment of the present application can also be a circuit or any other device that can realize a storage function, for storing program instructions and / or data.

[0110] By programming the processor 401, the code corresponding to the data recommendation method described in the above embodiment can be fixed into the chip, so that the chip can execute the code when running. Figure 1 The steps of the data recommendation method of the embodiment shown are as follows: How to design and program the processor 401 is a technique well known to those skilled in the art and will not be described in detail here.

[0111] Based on the same inventive concept, an embodiment of the present application further provides a storage medium, which stores computer instructions. When the computer instructions are executed on a computer, the computer executes the data recommendation method discussed above.

[0112] In some possible implementations, various aspects of the data recommendation method provided by the present application may also be implemented in the form of a program product, which includes program code. When the program product is run on an apparatus, the program code is used to enable the control device to execute the steps of the data recommendation method according to various exemplary implementations of the present application described above in this specification.

[0113] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0114] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0115] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0116] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0117] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A data recommendation method, characterized in that: include: Acquire current user behavior data, and select, from a feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with a historical user behavior feature is less than a set correlation; Based on the difference user behavior characteristics and the historical user behavior characteristics, a preset user interest prediction model is trained; The user interest prediction model is used to output recommendation results corresponding to the user's interest preferences.

2. The method according to claim 1, characterized in that The obtaining of current user behavior data includes: Obtain initial user behavior data; Performing data cleaning and denoising on the initial user behavior data to obtain target user behavior data; The target user behavior data is compressed and desensitized to obtain the current user behavior data.

3. The method according to claim 1, characterized in that The selecting, from the feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with the historical user behavior feature is less than a set correlation, includes: Based on a deep learning network, extract a feature set corresponding to the current user behavior data; Calculating the correlation between each feature in the feature set and the historical user behavior feature; Based on the correlation between each feature and the historical user behavior feature, each feature is sorted and the top n features are selected; The n features are used as the difference user behavior features, where n is an integer greater than 0.

4. The method according to claim 3, characterized in that The taking the n features as the difference user behavior features includes: Combining the n features according to their relevance to obtain m new features; The m newly added features and the n features are used as the difference user behavior features, where m is an integer greater than 0.

5. The method according to claim 1, characterized in that After outputting the recommendation results corresponding to the user's interest preferences through the user interest prediction model, the method further includes: Displaying the recommendation results on the homepage according to the set recommendation priority; Based on the user's homepage operations on the recommended homepage, the layout, theme or style of the recommended homepage is adjusted, and according to the feedback data returned by the user for the recommendation results, the recommendation strategy of the recommended homepage is adjusted in combination with reinforcement learning or online learning algorithms.

6. A data recommendation device, characterized in that: include: A data acquisition module, used to acquire current user behavior data, and select, from a feature set corresponding to the current user behavior data, a difference user behavior feature whose correlation with a historical user behavior feature is less than a set correlation; A model training module, used for training a preset user interest prediction model based on the differential user behavior characteristics and the historical user behavior characteristics; The result generation module is used to output the recommendation results corresponding to the user's interest preferences through the user interest prediction model.

7. The device according to claim 6, characterized in that The data acquisition module is specifically used for: Based on a deep learning network, extract a feature set corresponding to the current user behavior data; Calculating the correlation between each feature in the feature set and the historical user behavior feature; Based on the correlation between each feature and the historical user behavior feature, each feature is sorted and the top n features are selected; The n features are used as the difference user behavior features, where n is an integer greater than 0.

8. The device according to claim 6, characterized in that The device further includes a recommendation display module, which is specifically used for: Displaying the recommendation results on the homepage according to the set recommendation priority; Based on the user's homepage operations on the recommended homepage, the layout, theme or style of the recommended homepage is adjusted, and according to the feedback data returned by the user for the recommendation results, the recommendation strategy of the recommended homepage is adjusted in combination with reinforcement learning or online learning algorithms.

9. An electronic device, characterized in that: include: Memory, used to store computer programs; A processor, configured to implement the method steps of any one of claims 1 to 5 when executing the computer program stored in the memory.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method steps described in any one of claims 1 to 5 are implemented.