An information recommendation method, device, equipment and computer readable storage medium

Abstract

The application provides an information recommendation method and device, equipment and a computer readable storage medium. The method comprises the following steps: obtaining search behavior data of a target user; determining a material set corresponding to the search behavior according to the search behavior data; determining the type of operation behavior of the target user on the material corresponding to the search behavior in the material set; determining a subset of recommended materials corresponding to each type of operation behavior according to the recommendation strategy corresponding to different types of operation behavior; obtaining a set of recommended materials according to the priority of the subset of recommended materials; and recommending materials to the target user according to the set of recommended materials. Through the application, the information recommended for the target user is more in line with the interest points of the target user, and the accuracy of information recommendation can be improved.

Description

Technical Field

[0001] This application relates to computer processing techniques, and more particularly to an information recommendation method, apparatus, device, and computer-readable storage medium. Background Technology

[0002] Recommendation systems aim to increase user growth by recommending materials that users are interested in. Recommendation systems are divided into two stages: recall and ranking. Current recommendation recall algorithms mostly use click data as the basis for calculation. However, the main source of click data is the data that the platform recommends to users based on calculations. Users may not be interested in the pushed content, resulting in poor recommendation accuracy. Summary of the Invention

[0003] This application provides an information recommendation method, apparatus, and computer-readable storage medium, which can analyze users' interests based on their proactive behavior and improve the accuracy of information recommendation.

[0004] The technical solution of this application is implemented as follows:

[0005] This application provides an information recommendation method, comprising: obtaining search behavior data of a target user, wherein the search behavior data is used to indicate the target user's search behavior; determining a set of materials corresponding to the search behavior based on the search behavior data; determining the type of operation behavior of the target user on the materials corresponding to the search behavior in the set of materials; determining a subset of materials to be recommended for each type of operation behavior according to the recommendation strategy corresponding to each type of operation behavior; obtaining a set of materials to be recommended based on the priority of the subset of materials to be recommended; and recommending materials to the target user based on the set of materials to be recommended.

[0006] In some possible implementations, the target user's action on the materials corresponding to the search behavior is used to indicate the target user's level of interest in the materials corresponding to the search behavior; the recommendation strategy corresponding to different types of action behaviors determines the subset of materials to be recommended for each type of action behavior, including: determining a first subset of materials to be recommended based on the content data of the search behavior, based on the target user's first action on the materials corresponding to the search behavior; determining a second subset of materials to be recommended based on the content data of the materials corresponding to the search behavior, based on the target user's second action on the materials corresponding to the search behavior; wherein, the materials corresponding to the search behavior are obtained from a preset material library based on the content data of the search behavior; the target user's level of interest in the materials corresponding to the search behavior, indicated by the second action behavior, is higher than that of the first action behavior.

[0007] In some possible implementations, based on the first operational behavior of the target user towards the materials corresponding to the search behavior, a first subset of materials to be recommended is determined according to the content data of the search behavior. This includes: obtaining a first set of the top N most popular content data with the highest search volume from the content data of the preset user's search behavior, where N is an integer greater than 0; obtaining a second set of content data of the target user's search behavior; taking the intersection of the first set and the second set to obtain the popular content data of the target user's search behavior; calculating a first similarity between the popular content data and materials in a preset material library based on a similarity algorithm model; and determining the first subset of materials to be recommended from the preset material library based on the first similarity.

[0008] In some possible implementations, based on the target user's second action on the materials corresponding to the search behavior, a second subset of materials to be recommended is determined according to the content data of the materials corresponding to the search behavior, including: obtaining the content data of the materials corresponding to the search behavior; calculating the second similarity between the content data of the materials corresponding to the search behavior and the materials in a preset material library based on a similarity algorithm model; and determining the second subset of materials to be recommended from the preset material library based on the second similarity.

[0009] In some possible implementations, the set of materials to be recommended is obtained according to the priority of the subset of materials to be recommended, including: determining a second subset of materials to be recommended as the set of materials to be recommended; wherein the second subset of materials to be recommended is the subset of materials to be recommended with the highest priority.

[0010] In some possible implementations, obtaining a set of materials to be recommended based on the priority of the subset of materials to be recommended includes: determining whether the quantity of materials in the second subset of materials to be recommended is greater than or equal to a first threshold; when the quantity of materials in the second subset of materials to be recommended is greater than or equal to the first threshold, selecting materials from the second subset of materials to be recommended whose quantity meets the first threshold as the set of materials to be recommended; when the quantity of materials in the second subset of materials to be recommended is less than the first threshold, merging the first subset of materials to be recommended into the second subset of materials to be recommended, and selecting materials from the merged second subset of materials to be recommended whose quantity meets the first threshold as the set of materials to be recommended; wherein, the priority of the first subset of materials to be recommended is the next level below the priority of the second subset of materials to be recommended.

[0011] In some possible implementations, before determining the subset of materials to be recommended for each type of operation according to the recommendation strategy corresponding to each type of operation, the method further includes: performing data cleaning on the content data of the search behavior and / or the content data of the materials corresponding to the search behavior, so as to filter out invalid content in the content data of the search behavior and / or the content data of the materials corresponding to the search behavior.

[0012] Secondly, this application provides an information recommendation device, comprising: a first obtaining module, configured to obtain search behavior data of a target user, the search behavior data being used to indicate the target user's search behavior; a first determining module, configured to determine a set of materials corresponding to the search behavior based on the search behavior data; a second determining module, configured to determine the type of operation behavior of the target user on the materials corresponding to the search behavior in the set of materials; a third determining module, configured to determine a subset of materials to be recommended for each type of operation behavior according to the recommendation strategy corresponding to each type of operation behavior; a processing module, configured to obtain a set of materials to be recommended based on the priority of the subset of materials to be recommended; and a recommending module, configured to recommend materials to the target user based on the set of materials to be recommended.

[0013] In some possible implementations, the target user's operation on the material corresponding to the search behavior is used to indicate the target user's level of interest in the material corresponding to the search behavior; the search sub-behavior is used to indicate the target user's level of interest in the material corresponding to the search behavior; the third determining module is further used to determine a second subset of materials to be recommended based on the target user's second operation on the material corresponding to the search behavior and the content data of the material corresponding to the search behavior; wherein, the material corresponding to the search behavior is obtained from a preset material library based on the content data of the search behavior; the target user's level of interest in the material corresponding to the search behavior indicated by the second operation behavior is higher than that of the first operation behavior.

[0014] In some possible implementations, the third determining module is further configured to: obtain a first set of the top N most popular content data with the highest search volume from the content data of the preset user's search behavior; where N is an integer greater than 0; obtain a second set of content data of the target user's search behavior; take the intersection of the first set and the second set to obtain the popular content data of the target user's search behavior; calculate the first similarity between the popular content data of the target user's search behavior and the materials in the preset material library based on a similarity algorithm model; and determine a first subset of materials to be recommended from the preset material library based on the first similarity.

[0015] In some possible implementations, the third determining module is further configured to acquire content data of materials corresponding to the search behavior; calculate a second similarity between the content data of materials corresponding to the search behavior and materials in a preset material library based on a similarity algorithm model; and determine a second subset of materials to be recommended from the preset material library based on the second similarity.

[0016] In some possible implementations, the processing module is further configured to determine the second subset of materials to be recommended as the set of materials to be recommended; wherein the second subset of materials to be recommended is the subset of materials to be recommended with the highest priority.

[0017] In some possible implementations, the processing module is further configured to determine whether the quantity of materials in the second subset of materials to be recommended is greater than or equal to a first threshold; when the quantity of materials in the second subset of materials to be recommended is greater than or equal to the first threshold, select materials whose quantity meets the first threshold from the second subset of materials to be recommended as a set of materials to be recommended; when the quantity of materials in the second subset of materials to be recommended is less than the first threshold, merge the first subset of materials to be recommended into the second subset of materials to be recommended, and select materials whose quantity meets the first threshold from the merged second subset of materials to be recommended as a set of materials to be recommended; wherein, the priority of the first subset of materials to be recommended is the next level below the priority of the second subset of materials to be recommended.

[0018] In some possible implementations, the device further includes a data cleaning module, used to clean the content data of the search behavior and / or the content data of the material corresponding to the search behavior, so as to filter out invalid content in the content data of the search behavior and / or the content data of the material corresponding to the search behavior.

[0019] Thirdly, this application provides an information recommendation device, comprising:

[0020] Memory, used to store executable instructions;

[0021] A processor, when executing executable instructions stored in memory, implements the method provided in the first aspect of this application.

[0022] Fourthly, this application provides a computer-readable storage medium storing executable instructions for implementing the method provided in the first aspect of this application when executed by a processor.

[0023] The advantages of the technical solution provided in this application compared with the prior art are as follows:

[0024] In this application, based on the different types of operational behaviors of the target user towards the materials corresponding to the search behavior, different subsets of materials to be recommended are obtained for each type. The set of materials to be recommended is then obtained according to the priority of the subsets of materials to be recommended and recommended to the target user. In this way, the recommended materials determined by the target user's active search behavior are more in line with the target user's interests. At the same time, the hierarchical processing based on different types of operational behaviors can reflect the different levels of interest of the target user according to different operational behaviors, thereby ensuring that the recommended information is more likely to be high-quality information that matches the target user's interests and improving the accuracy of information recommendation. Attached Figure Description

[0025] Figure 1 This is a flowchart illustrating an information recommendation method in related technologies;

[0026] Figure 2This is a schematic diagram of the architecture of the information recommendation system provided in the embodiments of this application;

[0027] Figure 3 This is a schematic diagram of the server structure provided in an embodiment of this application;

[0028] Figure 4 This is a flowchart illustrating an information recommendation method provided in an embodiment of this application;

[0029] Figure 5 This is an optional flowchart illustrating the information recommendation method provided in the embodiments of this application;

[0030] Figure 6 This is an optional flowchart illustrating the information recommendation method provided in the embodiments of this application;

[0031] Figure 7 This is an optional flowchart illustrating the information recommendation method provided in the embodiments of this application;

[0032] Figure 8 This is a flowchart illustrating the application of the information recommendation method provided in this application to a social application software.

[0033] Figure 9 This is a schematic diagram of a computer device structure provided in an embodiment of this application. Detailed Implementation

[0034] To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings. The described embodiments should not be regarded as limitations on this application. All other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0035] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0036] In the following description, the terms "first / second / third" are used merely to distinguish similar objects and do not represent a specific ordering of objects. It is understood that "first / second / third" may be interchanged in a specific order or sequence where permitted, so that the embodiments of this application described herein can be implemented in an order other than that illustrated or described herein.

[0037] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0038] To illustrate the technical solution described in this application, specific embodiments are provided below.

[0039] With the rapid development of technology, especially the widespread use of mobile communication networks and mobile terminals, various materials and content exist on the Internet. Users can also produce and create materials and content anytime and anywhere. Different manufacturers provide platforms, which on the one hand collect some of the materials and content produced by users, and on the other hand, filter high-quality materials and content through different methods, and push corresponding materials and content to users based on different user interests, so as to increase the probability of users clicking to view the pushed materials and content data.

[0040] Various online platforms typically store large amounts of content, forming a content database. This content data can include multimedia data in various formats, such as video, audio, image, and text data. The format varies depending on the business scenario, for example, live broadcasts, short videos, songs, comics, audiobooks, novels, news, articles, and food platforms. Platforms can proactively publish this multimedia data; for example, they can purchase various multimedia data from copyright holders for display. Alternatively, registered users on the platform can generate various types of multimedia data using electronic devices and upload them to the platform. Other users can then search and browse this multimedia data using electronic devices, including but not limited to laptops, tablets, desktop computers, and mobile terminals.

[0041] Different business scenarios can involve various forms of multimedia data. For example, in a live streaming scenario, users can use cameras to capture video data and perform operations such as beautification and live interaction to generate live programs. In a short video scenario, users can use cameras to capture video data and perform operations such as beautification, editing, and adding special effects to generate short videos. In an information sharing and communication scenario, users can upload edited text, images, videos, etc., to generate blog posts. This multimedia data can be published on corresponding platforms, and other users can access this data on the platforms using smart devices such as terminals, or share this data content.

[0042] Recommendation systems emerged to more accurately filter content from massive amounts of data and recommend it to users. By showing users materials they are likely to be interested in, recommendation systems can attract new users and increase monthly active users, thereby achieving user growth. Essentially, it recommends high-click-rate materials to users and is widely used in e-commerce, search, and advertising to recommend personalized content. For example, in advertising scenarios, personalized recommendation systems can push advertising materials to users based on their characteristics and preferences. If the user ultimately clicks and converts, the advertising push is considered successful; otherwise, it fails.

[0043] Recommendation systems typically include two phases: recall and ranking. Current recall algorithms usually use user click data as the basis for calculations. Based on whether a user clicks on a recommendation, they determine whether the user has some interest in that recommendation and then recommend other similar data to the user. See also... Figure 1 As shown, Figure 1 This is a flowchart illustrating an information recommendation method in related technologies. The information recommendation system first obtains the user's click data (S1) and obtains preset interest rules and / or calculation models (S2). Then, using the preset interest rules and / or calculation models, it determines materials from the material database that are similar to the content of the user's click data and recommends them to the user (S3).

[0044] However, the primary goal of a recommendation system is to more accurately capture user interests and meet their personalized needs. The recommendation information data that users click on is data that the platform recommends to users after calculation, such as users clicking on or reading recommendation information. This is information that users passively acquire. If we rely solely on the user's interest points determined by their behavior in passively acquiring information, it does not actually reflect the user's subjective interests. Therefore, it is possible for the recommendation to fail because the recommended information is not information that the user is interested in.

[0045] To address the aforementioned issues, embodiments of this application provide an information recommendation method, apparatus, device, and computer-readable storage medium, wherein the determined recommended information is high-quality information that better matches the interests of the target user, thereby improving the effectiveness of information recommendation.

[0046] The information recommendation method provided in this application can be implemented by various electronic devices. For example, it can be implemented by a server alone, or by a terminal and a server working together. For example, the server can execute the live streaming recommendation method described below on its own, or the terminal can send a recommendation request message to the server, and the server can execute the information recommendation method based on the received recommendation request message.

[0047] The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, and big data and artificial intelligence platforms. The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, and this application embodiment does not impose specific limitations in this regard.

[0048] See Figure 2 , Figure 2 This is a schematic diagram of the architecture of the information recommendation system 200 provided in the embodiments of this application. In this system, terminals (terminals 201-1 and 201-2 are shown as examples) are connected to the server 203 through a network 202. The network 202 can be a wide area network or a local area network, or a combination of both.

[0049] In some embodiments, the terminal or server 200 can implement the information recommendation method provided in this application embodiment by running a computer program. For example, the computer program can be a native program or software module in an operating system; for instance, it can be a native application (APP), that is, a program that needs to be installed in the operating system to run, such as a live streaming APP; it can also be a live streaming mini-program, that is, a live streaming program that only needs to be downloaded to a browser environment to run; or it can be a live streaming mini-program that can be embedded in any APP. In short, the above-mentioned computer program can be any form of application, module or plugin, and this application embodiment does not specifically limit it.

[0050] Here, we will use terminal 201-1 as the target user for illustration. Terminal 201-1 displays the APP interface on the current interface 210-1. For example, terminal 201-1 can also display recommended materials shown to the target user based on the target user's search behavior on the current interface. Server 200 can receive the content data of the search behavior sent by terminal 201-1, where the search behavior includes multiple sub-behaviors. Then, the server breaks down the sub-behaviors of the target user's search behavior and recommends the set of recommended materials to the target user according to the recommendation strategy corresponding to different search behaviors.

[0051] See Figure 3 , Figure 3 This is a schematic diagram of the server structure provided in an embodiment of this application. Figure 3The server 203 shown may include at least one processor 310, memory 320, at least one network interface 330, and user interface 340. The various components in device 300 are coupled together via a bus system 350. It is understood that the bus system 350 is used to implement communication between these components. In addition to a data bus, the bus system 350 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in… Figure 3 The general labeled all buses as Bus System 350.

[0052] The processor 310 can be an integrated circuit chip with signal processing capabilities, such as a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0053] User interface 340 includes one or more output devices 341 that enable the presentation of media content, including one or more speakers and / or one or more visual displays. User interface 304 also includes one or more input devices 342, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0054] The memory 320 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state storage, hard disk drives, optical disk drives, etc. The memory 320 may optionally include one or more storage devices physically located away from the processor 310.

[0055] The memory 320 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), and the volatile memory may be random access memory (RAM). The memory 320 described in this application embodiment is intended to include any suitable type of memory.

[0056] In some embodiments, memory 320 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, as illustrated below.

[0057] Operating system 321 includes system programs for handling various basic system services and performing hardware-related tasks, such as the framework layer, core library layer, driver layer, etc., for implementing various basic business functions and handling hardware-based tasks;

[0058] The network communication module 322 is used to reach other computing devices via one or more (wired or wireless) network interfaces 330, exemplary network interfaces 330 including: Bluetooth, WiFi, and Universal Serial Bus (USB), etc.

[0059] Presentation module 323 is configured to enable the presentation of information (e.g., a user interface for operating peripheral devices and displaying content and information) via one or more output devices 342 (e.g., a display screen, a speaker, etc.) associated with user interface 340;

[0060] The input processing module 324 is used to detect and translate one or more user inputs or interactions from one or more input devices 342.

[0061] In some embodiments, the user recall device provided in this application can be implemented in software. Figure 3 A user recall device 325 stored in memory 320 is shown. This device can be software in the form of programs and plug-ins, and includes the following software modules: a first obtaining module 3251, a first determining module 3252, a second determining module 3253, a third determining module 3254, a processing module 3255, and a recommendation module 3256. These modules are logically connected and can therefore be arbitrarily combined or further divided according to their implemented functions. The functions of each module will be described below.

[0062] In other embodiments, the user recall device provided in this application can be implemented in hardware. As an example, the user recall device provided in this application can be a processor in the form of a hardware decoding processor, which is programmed to execute the user recall method provided in this application. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0063] In other embodiments, the information recommendation device provided in this application can be implemented in hardware. As an example, the information recommendation device provided in this application can be a processor in the form of a hardware decoding processor, which is programmed to execute the information recommendation method provided in this application. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0064] The following will describe the information recommendation method provided in this application embodiment by referring to exemplary applications and implementations of the information recommendation device provided in the embodiments of this application.

[0065] See Figure 4 , Figure 4 This is a flowchart illustrating an information recommendation method provided in an embodiment of this application. The following will be combined with... Figure 4 The steps shown are explained.

[0066] S401, Obtain search behavior data of the target user;

[0067] Among them, search behavior data is used to indicate the search behavior of target users.

[0068] In some embodiments, the target user's search behavior data may include the target user's search terms, or it may also include data on the time and location of the target user's search behavior. The search terms are data actively entered by the target user. The number of search terms can be one or more, such as a single character, multiple words, or even a long sentence. The content of the search terms can be text with direct sentence meaning, or it can be meaningless symbols, such as some internet slang. Furthermore, the search terms can be of any language type, such as English, Chinese, or a combination of multiple languages.

[0069] In some embodiments, search behavior data may not be limited to text information; the target user's search behavior data may also be image information. For example, for platforms that support image retrieval, the target user can upload the image information they want to search for to find content related to the image information content. In this case, the information recommendation device can obtain the image information, determine the related recommended materials based on the content of the image information, and recommend them to the target user.

[0070] S402, determine the set of materials corresponding to the search behavior based on the search behavior data.

[0071] Understandably, the set of materials corresponding to a search behavior can be a set of one or more related materials determined from the material library based on the search terms, images, and other data mentioned above.

[0072] For example, after a target user enters a search term through a terminal, the information recommendation device retrieves materials containing that search term from the material database and sends them to the terminal. The terminal then displays the received materials through a display interface. The received materials can be of any type, such as text, images, or videos.

[0073] S403, Determine the type of operation behavior of the target user on the material corresponding to the search behavior in the material set.

[0074] In some embodiments, the target user's operation on the material corresponding to the search behavior in the material set is a further operation on the searched material after the target user has found the corresponding material through the search behavior.

[0075] For example, after the information recommendation device retrieves one or more materials, including text, images, and videos, based on the aforementioned search terms and sends them to the terminal, the target user can view these materials through the terminal's display interface. Further actions taken by the target user based on the retrieved materials may include: clicking on one or more materials to view them, liking, saving, forwarding, commenting on the content of one or more materials, etc.

[0076] Understandably, recommendation systems typically analyze user behavior data to understand user interests, such as preferred data and types, like image styles and genres, and video styles. The different actions a target user takes in response to materials found based on search terms can also reflect their level of interest in the content. For example, if a terminal receives multiple materials, and the target user clicks to view one or more of them, their interest in the viewed materials is likely higher than in the unviewed materials. Similarly, among multiple viewed materials, the probability that the target user is more interested in the saved materials than in the unsaved ones is also higher. Furthermore, other user actions that can indicate different levels of interest can also be included, such as the varying lengths of time a target user spends viewing multiple materials.

[0077] S404, according to the recommendation strategy corresponding to different types of operation behaviors, determine the corresponding subset of materials to be recommended for each type of operation behavior;

[0078] In some embodiments, the target user's actions may correspond to only one material.

[0079] For example, for a given piece of material, such as video content, the user action could include whether the target user clicked to watch the video or not. In this case, based on the action of clicking to watch, determining the subset of materials to be recommended can be done using the video's content information as a search marker, such as explanatory videos or photo albums of the people featured in the video. Conversely, for the action of not clicking to watch, determining the subset of materials to be recommended can be done using data such as the search terms used by the target user to find the video as search markers.

[0080] In other embodiments, each operation may correspond to multiple materials.

[0081] For example, again taking video content as the material corresponding to the user's action, the target user's action may include clicking to watch these videos, or not clicking to watch these videos. Furthermore, since the duration of multiple videos varies, the target user's action may also include watching the entire video, or not watching the entire video content. Alternatively, it could include the duration of each video watched by the target user and whether they liked, shared, or commented on the video. First, for the action of not clicking to watch, determining the subset of materials to be recommended to the target user can be based on the search terms used by the target user to find these videos as search markers. For the action of the target user clicking to watch videos, since the duration of each video watched by the target user varies, or some videos were only clicked to view without being watched in their entirety, the subset of materials to be recommended to the target user can also be determined primarily based on the video content that the target user watched in its entirety.

[0082] Understandably, the type of actions a target user takes in relation to materials based on their search behavior can, to some extent, reflect their interest in those materials. Therefore, in this embodiment, the target user's actions on materials related to their search behavior can be used to indicate their level of interest. Based on this interest level, different strategies are used to determine recommended materials for the target user. The resulting subset of recommended materials better matches the target user's interests, thereby improving the recommendation effectiveness.

[0083] For example, see Figure 5 As shown, Figure 5 An optional flowchart of the information recommendation method provided in this application embodiment is based on... Figure 5 Step S403 above may include:

[0084] S501, based on the first operation behavior of the target user for the materials corresponding to the search behavior, determine the first subset of materials to be recommended according to the content data of the search behavior;

[0085] Among them, the content data of search behavior can be the data entered by the target user when obtaining the corresponding material based on the search behavior, such as the search terms and images mentioned above.

[0086] S502, based on the target user's second operation behavior for the materials corresponding to the search behavior, determine the second recommended material subset according to the content data of the materials corresponding to the search behavior.

[0087] Among them, the target user indicated by the second operation behavior is more interested in the materials corresponding to the search behavior than the first operation behavior.

[0088] It is understood that the first and second operation behaviors described herein are only used to distinguish different sub-operation behaviors of the target user toward the material. Based on multiple types of operation behaviors, the embodiments of this application may also include multiple sub-operations to indicate different degrees of interest of the target user toward the material corresponding to the search behavior.

[0089] For example, the second action may refer to the target user clicking on one or more materials after the search action, while the first action may refer to the materials that were not clicked after the search action.

[0090] Understandably, for target users, the materials clicked after a search are likely to elicit a higher level of interest compared to materials not clicked after a search. Therefore, this application embodiment categorizes target users' different types of operational behaviors towards materials corresponding to their search behavior, representing the target user's level of interest in different materials. Based on this level of interest, different strategies are used to determine recommended materials for the target user, making them more aligned with their interests and thus improving the recommendation effectiveness.

[0091] The materials corresponding to the search behavior are obtained from a preset material library based on the content data corresponding to the search behavior. For example, the content data of the search behavior is the search term, and the materials corresponding to the search behavior can be materials that the information recommendation device finds from the material library based on the search term, containing all or part of the content of the search term. The content data of the materials corresponding to the search behavior indicates the data of the material content. For example, if the material corresponding to the search behavior is a video file, the content data of the material can include data information such as the duration, style type, and video content of the video file.

[0092] In one embodiment, see Figure 6 As shown, Figure 6 This is an optional flowchart illustrating the information recommendation method provided in an embodiment of this application, based on... Figure 6 The above step S501 may include:

[0093] S601, Obtain the first set of the top N most popular content data with the highest search volume from the content data of the preset user's search behavior;

[0094] It should be noted that preset users can be any set of users to be recommended in the recommendation system, that is, each target user. For example, if the target user is a registered user on a social media platform, the preset users could refer to all registered users on that platform, or the preset users could be a subset of registered users on the platform determined by the recommendation system based on certain conditions.

[0095] The content data for the preset user's search behavior consists of the input data from the preset user's search behavior, such as search terms. The top N most searched content data consists of the top N most frequently repeated input data from all search behaviors of the preset user. For example, when the preset user is all registered users of the platform, the content data of the preset user's search behavior can be the content data of the search behavior of each registered user obtained by the information recommendation device, that is, the set of input data (search terms) from the search behavior of each registered user on the platform. In this case, the top N most searched content data consists of the most frequently repeated search terms from the search terms of all registered users on the platform.

[0096] The value of N can be obtained based on data analysis. For example, when the user group is large, the total number of search terms will inevitably be large, and the required computing resources will also grow exponentially. In addition, all search terms will inevitably contain a lot of invalid content such as long-tail keywords, and some search terms with low search volume will also have fewer searchers and a smaller impact. Therefore, choosing a reasonable threshold will ensure the maximum computing effect while saving computing resources.

[0097] S602, Obtain the second set of content data related to the target user's search behavior;

[0098] In some embodiments, the second set of content data on the target user's search behavior may be data on the target user's search behavior within a predetermined time window.

[0099] Understandably, the longer the time span of search behavior, the more content data related to the target user's search behavior may be available. However, this data may not reflect the target user's level of interest as much. For example, a target user might have searched for a TV series a month ago, but they may have already watched it within the past month. Therefore, this type of search behavior data is not very useful for determining the target user's level of interest and may even affect the final recommendation results. Therefore, obtaining content data on search behavior within a predetermined time window (e.g., the most recent day or three days) can best fit the target user's current interests and improve the accuracy of information recommendation.

[0100] S603, take the intersection of the first set and the second set to obtain the popular content data of the target user's search behavior;

[0101] Understandably, taking the example of a target user obtaining relevant video content by entering search terms, if the target user only searches but doesn't click, it likely indicates that the target user isn't very interested in the video content found based on that search term, or that some target users may not have a specific purpose for being interested in that content. Therefore, this embodiment of the application takes the intersection of a first set of the target user's search term records and the popular search terms of the entire user group, and tries to fit the content that the target user is interested in as much as possible with the interest probability of the majority of people. This can increase the probability that the determined set of recommended materials is content that the target user is interested in, even when there is limited basic data on the target user's interest points, thus helping to improve the recommendation effect.

[0102] S604, Based on popular content data, calculate the first similarity between the popular content data of the target user's search behavior and the materials in the preset material library based on the similarity algorithm model;

[0103] The similarity algorithm model can be any existing model, such as a system filtering algorithm. Alternatively, it can be trained based on the target user's historical data. This application does not impose specific limitations on the selection of the similarity algorithm model.

[0104] In some embodiments, the information recommendation device can vectorize each of the N popular search terms using a word vectorization algorithm, extract the meaning of each search term within the complete search term using a semantic analysis model, and perform data cleaning to remove meaningless modal particles, etc., to obtain the vectorized representation of each popular search term. Then, by taking the intersection of the second set of the target user's search terms with the N popular search terms, the vectorized representation of each target user's popular search terms can be obtained. The semantic analysis model can be an existing trained model, such as the Bidirectional Encoder Representations from Transformers (BERT) algorithm.

[0105] S605, determine the first subset of materials to be recommended from the preset material library based on the first similarity.

[0106] In some embodiments, the information recommendation device can obtain a first material to be recommended from a preset material library by setting a similarity threshold.

[0107] For example, the information recommendation device identifies all materials in the preset material library that meet the similarity threshold with the target user's popular search terms as the first subset of materials to be recommended.

[0108] In other embodiments, the information recommendation device may also obtain a first subset of materials to be recommended from a preset material library by setting a first threshold for the number of materials to be recommended.

[0109] For example, the quantity threshold can be 50. In this case, the information recommendation device can sort the materials in the preset material database based on the similarity value and select the top 50 materials as the first set of materials to be recommended.

[0110] In some embodiments, see Figure 7 As shown, Figure 7 This is an optional flowchart illustrating the information recommendation method provided in an embodiment of this application, based on... Figure 7 Step S502 above may include:

[0111] S701, Obtain the content data of the material corresponding to the search behavior;

[0112] Understandably, the search behavior corresponds to materials recommended by the information recommendation device. Based on search behavior data, such as search terms, the device searches the material library for materials containing all or part of the search terms. When a target user clicks on one or more materials to view them after searching, it indicates that these materials may be of interest to the target user. Therefore, materials of a certain type can be recommended to the target user based on the content corresponding to these materials.

[0113] S702, Based on the content data of the material corresponding to the search behavior, calculate the second similarity between the content data of the material corresponding to the search behavior and the materials in the preset material library based on the similarity algorithm model;

[0114] The similarity algorithm can be any existing algorithm model, such as the same similarity algorithm used to calculate the similarity of search terms and materials as described above. This application does not impose specific limitations on the selection of the similarity algorithm.

[0115] S703, determine a second subset of materials to be recommended from a preset material library based on the second similarity.

[0116] In some embodiments, the information recommendation device can obtain a second material to be recommended from a preset material library by setting a similarity threshold.

[0117] For example, all materials in the preset material library whose content data matches the search behavior's content data and meet the similarity threshold are identified as the second set of materials to be recommended.

[0118] In other embodiments, the server can also obtain a second subset of materials to be recommended from a preset material library by setting a threshold for the quantity of the second set of materials to be recommended.

[0119] For example, the threshold number could be 50. In this case, the materials in the preset material database can be sorted based on similarity, and the top 50 materials can be selected as the second set of materials to be recommended.

[0120] S405, Based on the priority of the subset of materials to be recommended, obtain the set of materials to be recommended;

[0121] It should be noted that the priority of the subset of recommended materials is actually the priority of the corresponding operation. For example, the target user is more interested in the subset of recommended materials corresponding to the operation of clicking on the search than in the subset of recommended materials corresponding to the operation of not clicking on the search. Therefore, the subset of recommended materials corresponding to the operation of clicking on the search has a higher priority than the subset of recommended materials corresponding to the operation of not clicking on the search.

[0122] Understandably, the first subset to be recommended can be determined based on search behavior data, such as search terms. Each target user can have one or more search terms. Therefore, for any given target user, a first subset of materials to be recommended can be determined based on search terms. For target users who have performed the aforementioned second action, the corresponding second subset of materials to be recommended obtained based on that second action shows a higher level of interest. Therefore, the second subset of materials to be recommended can be prioritized as the set of materials to be recommended to the target user.

[0123] Based on this, in some embodiments, obtaining the set of materials to be recommended according to the priority of the subset of materials to be recommended may include: determining the second subset of materials to be recommended as the set of materials to be recommended.

[0124] Among them, the second subset of materials to be recommended is the subset of materials to be recommended with the highest priority.

[0125] In addition, since there may be a situation where the number of materials in the second subset of materials to be recommended is small, the first subset of materials to be recommended and the second subset of materials to be recommended can be merged, and the merged first subset of materials to be recommended and the second subset of materials to be recommended can be used together as a set of materials to be recommended to the target user.

[0126] Based on this, in some embodiments, obtaining the set of materials to be recommended according to the priority of the subset of materials to be recommended may further include the following steps:

[0127] Step 1: Determine whether the number of materials in the second subset of materials to be recommended is greater than or equal to the first threshold.

[0128] Step 2: When the quantity of materials in the second subset of materials to be recommended is greater than or equal to the first threshold, select materials whose quantity satisfies the first threshold from the second subset of materials to be recommended as the set of materials to be recommended.

[0129] Step 3: When the quantity of materials in the second subset of materials to be recommended is less than the first threshold, the first subset of materials to be recommended is merged into the second subset of materials to be recommended. Materials whose quantity meets the first threshold are selected from the merged second subset of materials to be recommended as the set of materials to be recommended.

[0130] The priority of the first subset of materials to be recommended is the next level below the priority of the second subset of materials to be recommended.

[0131] In some embodiments, merging the first subset of materials to be recommended into the second subset of materials to be recommended may involve merging all materials in the first subset of materials to be recommended into the second subset of materials to be recommended, and then selecting materials whose quantity meets the first threshold from the merged second subset of materials to be recommended to the target user.

[0132] In other embodiments, merging the first subset of materials to be recommended into the second subset of materials to be recommended may involve first determining the difference between the quantity of materials in the second subset of materials to be recommended and a first threshold, and then selecting materials from the first subset of materials to be recommended whose quantity meets the difference and merging them into the second subset of materials to be recommended. For example, the materials in the first subset of materials to be recommended may be sorted based on a first similarity determined when determining the first subset of materials to be recommended, and then the material with the highest similarity may be selected to fill the difference.

[0133] For example, the first threshold is 30. There are only 20 materials in the current second set of materials to be recommended. Based on the calculated first similarity, the information recommendation device first sorts the materials in the first subset of materials to be recommended, then selects the 10 materials with the highest similarity and adds them to the second subset of materials to be recommended, and then determines the merged second subset of materials to be recommended as the set of materials to be recommended.

[0134] S406, based on the merged set of materials to be recommended, recommends materials to the target user.

[0135] Understandably, the second subset of materials to be recommended is the set of materials that the target user is most interested in. Therefore, when recommending materials to the target user, if the set of materials to be recommended is a combination of the first and second subsets, the materials in the second subset can be displayed first, placing the materials from the first subset after those from the second subset, and recommending the materials to the target user in order. In this way, when the target user sees the recommended materials on their device, they can prioritize viewing the materials that interest them most, improving the effectiveness of information recommendation.

[0136] In this embodiment, based on the different types of operation behaviors of the target user on the materials corresponding to the search behavior, different subsets of materials to be recommended are obtained for each type. The set of materials to be recommended is obtained according to the priority of the subsets of materials to be recommended and recommended to the target user. In this way, the recommended materials determined by the target user's active search behavior are more in line with the target user's interests. At the same time, the hierarchical processing based on different types of operation behaviors can reflect the different levels of interest of the target user according to different operation behaviors, thereby ensuring that the recommended information is more likely to be high-quality information that matches the target user's interests and improving the accuracy of information recommendation.

[0137] The following will describe an exemplary application of the embodiments of this application in a real-world application scenario.

[0138] When the above information recommendation method is applied to social media applications on a device, such as Weibo, the information recommended to the user can include blog posts, pictures, videos, etc., published by the blogger. The implementation process of applying the above information recommendation method to Weibo can be found in [link to relevant documentation]. Figure 8 , Figure 8 This is a flowchart illustrating the application of the information recommendation method provided in this application to a social networking application. The following will combine... Figure 8 The steps shown are explained.

[0139] S801, acquire the target user's search behavior data;

[0140] S802, determine the set of materials corresponding to the search behavior based on the search behavior data;

[0141] S803, determine the type of operation behavior of the target user on the material corresponding to the search behavior in the material set. When the type of operation behavior is the first operation behavior, execute steps S804 to S809; when the type of operation behavior is the second operation behavior, execute steps S810 to S814.

[0142] S804, obtain the first set of the top N most popular content data with the highest search volume from the content data of the preset user's search behavior;

[0143] S805, Obtain the second set of content data related to the target user's search behavior;

[0144] S806: Take the intersection of the first set and the second set to obtain the popular content data of the target user's search behavior;

[0145] S807, calculates the first similarity between the target user's popular content data and the materials in the preset material library based on the similarity algorithm model;

[0146] S808, Based on the first similarity, determine the first subset of materials to be recommended from the preset material library;

[0147] S809, Select materials whose quantity meets the first threshold from the first subset of materials to be recommended and determine them as the set of materials to be recommended;

[0148] S810, obtain the content data of the material corresponding to the search behavior;

[0149] S811, Based on the content data of the material corresponding to the search behavior, calculate the second similarity between the content data of the material corresponding to the search behavior and the materials in the preset material library based on the similarity algorithm model;

[0150] S812, determine a second subset of materials to be recommended from a preset material library based on the second similarity;

[0151] S813, determine whether the number of materials in the second subset of materials to be recommended is greater than or equal to the first threshold; if it is greater than or equal to the first threshold, proceed to step S814; if it is less than the first threshold, proceed to step S814.

[0152] S813, Select materials whose quantity meets the first threshold from the second subset of materials to be recommended as the set of materials to be recommended;

[0153] S814, merge the first subset of materials to be recommended into the second subset of materials to be recommended, and select materials whose quantity meets the first threshold from the merged second subset of materials to be recommended as the set of materials to be recommended;

[0154] S815 recommends materials to target users based on the set of materials to be recommended.

[0155] This completes the above information recommendation method.

[0156] The following description continues to illustrate the exemplary structure of the information recommendation device 325 provided in the embodiments of this application as a software module. In some embodiments, such as... Figure 3 As shown, the software modules stored in the information recommendation device 325 in the memory 320 may include: a first obtaining module 3251, used to obtain search behavior data of the target user, the search behavior data being used to indicate the target user's search behavior; a first determining module 3252, used to determine the material set corresponding to the search behavior based on the search behavior data; a second determining module 3253, used to determine the type of operation behavior of the target user on the materials corresponding to the search behavior in the material set; a third determining module 3254, used to determine the subset of materials to be recommended for each type of operation behavior according to the recommendation strategy corresponding to each type of operation behavior; a processing module 3255, used to obtain the set of materials to be recommended based on the priority of the subset of materials to be recommended; and a recommendation module 3256, used to recommend materials to the target user based on the set of materials to be recommended.

[0157] In some possible implementations, the target user's operation on the material corresponding to the search behavior is used to indicate the target user's level of interest in the material corresponding to the search behavior; the search sub-behavior is used to indicate the target user's level of interest in the material corresponding to the search behavior; the third determining module 3254 is further used to determine a second subset of materials to be recommended based on the target user's second operation on the material corresponding to the search behavior and the content data of the material corresponding to the search behavior; wherein, the material corresponding to the search behavior is obtained from a preset material library based on the content data of the search behavior; the target user's level of interest in the material corresponding to the search behavior indicated by the second operation behavior is higher than that of the first operation behavior.

[0158] In some possible implementations, the third determining module 3254 is further configured to: obtain a first set of the top N most popular content data with the highest search volume from the content data of the preset user's search behavior; where N is an integer greater than 0; obtain a second set of content data of the target user's search behavior; take the intersection of the first set and the second set to obtain the popular content data of the target user's search behavior; calculate the first similarity between the popular content data of the target user's search behavior and the materials in the preset material library based on a similarity algorithm model; and determine a first subset of materials to be recommended from the preset material library based on the first similarity.

[0159] In some possible implementations, the third determining module 3254 is further configured to acquire content data of materials corresponding to the search behavior; calculate a second similarity between the content data of materials corresponding to the search behavior and materials in a preset material library based on a similarity algorithm model; and determine a second subset of materials to be recommended from the preset material library based on the second similarity.

[0160] In some possible implementations, the processing module 3255 is further configured to determine the second subset of materials to be recommended as the set of materials to be recommended; wherein the second subset of materials to be recommended is the subset of materials to be recommended with the highest priority.

[0161] In some possible implementations, the processing module 3255 is further configured to determine whether the quantity of materials in the second subset of materials to be recommended is greater than or equal to a first threshold; when the quantity of materials in the second subset of materials to be recommended is greater than or equal to the first threshold, select materials whose quantity meets the first threshold from the second subset of materials to be recommended as a set of materials to be recommended; when the quantity of materials in the second subset of materials to be recommended is less than the first threshold, merge the first subset of materials to be recommended into the second subset of materials to be recommended, and select materials whose quantity meets the first threshold from the merged second subset of materials to be recommended as a set of materials to be recommended; wherein, the priority of the first subset of materials to be recommended is the next level below the priority of the second subset of materials to be recommended.

[0162] In some possible implementations, the information recommendation device further includes a data cleaning module, used to clean the content data of the search behavior and / or the content data of the materials corresponding to the search behavior, so as to filter out invalid content in the content data of the search behavior and / or the content data of the materials corresponding to the search behavior.

[0163] Figure 9 This is a schematic diagram of a computer device structure provided in an embodiment of this application. The computer device 910 can vary significantly due to different configurations or performance. It may include one or more central processing units (CPUs) 911 (e.g., one or more processors) and a memory 912, and one or more storage media 913 (e.g., one or more mass storage devices) for storing application programs 921 or data 922. The memory 912 and storage media 913 can be temporary or persistent storage. The program stored in the storage media 913 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the computer device. Furthermore, the CPU 911 may be configured to communicate with the storage media 913 and execute the series of instruction operations in the storage media 913 on the computer device 910.

[0164] The computer device 910 may also include one or more power supplies 914, one or more wired or wireless network interfaces 915, one or more input / output interfaces 916, and / or one or more operating systems 917, such as Windows Server™, Mac OS X™, Uni™, Linux™, FreeBSD™, etc.

[0165] This application provides a computer program product or computer program that includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the information recommendation method described in this application.

[0166] This application provides a computer-readable storage medium storing executable instructions. When these executable instructions are executed by a processor, they cause the processor to execute the information recommendation method provided in this application, for example... Figure 4 The information recommendation method is shown.

[0167] In some embodiments, the computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EP ROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or it may be a variety of devices including one or any combination of the above-mentioned memories.

[0168] In some embodiments, executable instructions may take the form of a program, software, software module, script, or code, written in any form of programming language (including compiled or interpreted languages, or declarative or procedural languages), and may be deployed in any form, including as a standalone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0169] As an example, executable instructions may, but do not necessarily, correspond to files in a file system. They may be stored as part of a file that holds other programs or data, for example, in one or more scripts in a Hyper Text Markup Language (HTML) document, in a single file dedicated to the program in question, or in multiple collaborating files (e.g., a file that stores one or more modules, subroutines, or code sections).

[0170] As an example, executable instructions can be deployed to execute on a single computing device, or on multiple computing devices located in one location, or on multiple computing devices distributed across multiple locations and interconnected via a communication network.

[0171] In summary, this application embodiment obtains different subsets of materials to be recommended based on the different types of operational behaviors of the target user towards the materials corresponding to the search behavior. The set of materials to be recommended is then obtained according to the priority of these subsets and recommended to the target user. Thus, the recommended materials determined through the target user's active search behavior are more aligned with the target user's interests. Furthermore, the hierarchical processing based on different types of operational behaviors reflects the target user's varying degrees of interest, thereby ensuring a higher probability that the recommended information is high-quality information that matches the target user's interests and improving the accuracy of information recommendation.

[0172] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Any modifications, equivalent substitutions, and improvements made within the spirit and scope of this application are included within the scope of protection of this application.

Claims

1. An information recommendation method, characterized in that, The method includes: Obtain search behavior data of the target user, the search behavior data being used to indicate the search behavior of the target user; The material set corresponding to the search behavior is determined based on the search behavior data; Determine the type of operation behavior of the target user on the materials corresponding to the search behavior in the material set; Based on the recommendation strategies corresponding to different types of operational behaviors, determine the subset of materials to be recommended for each type of operational behavior; Based on the priority of the subset of materials to be recommended, a set of materials to be recommended is obtained; Based on the set of materials to be recommended, materials are recommended to the target user; The target user's actions on the materials corresponding to the search behavior are used to indicate the target user's level of interest in the materials corresponding to the search behavior; The step of determining the subset of materials to be recommended for each type of operational behavior according to the recommendation strategy corresponding to different types of operational behaviors includes: Based on the first operation behavior of the target user towards the materials corresponding to the search behavior, a first subset of materials to be recommended is determined according to the content data of the search behavior. Based on the second operation behavior of the target user towards the materials corresponding to the search behavior, a second subset of materials to be recommended is determined according to the content data of the materials corresponding to the search behavior; The materials corresponding to the search behavior are obtained from a preset material library based on the content data of the search behavior; the target user indicated by the second operation behavior has a higher degree of interest in the materials corresponding to the search behavior than the target user indicated by the first operation behavior. The step of determining a first subset of materials to be recommended based on the first operation behavior of the target user for the materials corresponding to the search behavior, according to the content data of the search behavior, includes: Get the first set of the top N most popular content data with the highest search volume from the content data of the search behavior of the preset user, where N is an integer greater than 0; Obtain a second set of content data related to the search behavior of the target user; By taking the intersection of the first set and the second set, the popular content data of the target user's search behavior is obtained; Based on the popular content data, the first similarity between the popular content data of the target user's search behavior and the materials in the preset material library is calculated using a similarity algorithm model. The first subset of materials to be recommended is determined from the preset material library based on the first similarity.

2. The method according to claim 1, characterized in that, The second operation behavior based on the target user's search behavior for materials, determining a second subset of materials to be recommended based on the content data of the materials corresponding to the search behavior, includes: Obtain the content data of the material corresponding to the search behavior; Based on the content data of the material corresponding to the search behavior, a second similarity between the content data of the material corresponding to the search behavior and the materials in the preset material library is calculated using a similarity algorithm model; The second subset of materials to be recommended is determined from the preset material library based on the second similarity.

3. The method according to claim 1 or 2, characterized in that, The step of obtaining the set of materials to be recommended based on the priority of the subset of materials to be recommended includes: The second subset of materials to be recommended is determined as the set of materials to be recommended; wherein, the second subset of materials to be recommended is the subset of materials to be recommended with the highest priority.

4. The method according to claim 3, characterized in that, The step of obtaining the set of materials to be recommended based on the priority of the subset of materials to be recommended includes: Determine whether the number of materials in the second subset of materials to be recommended is greater than or equal to the first threshold; When the number of materials in the second subset of materials to be recommended is greater than or equal to the first threshold, materials whose number of materials in the second subset of materials to be recommended are selected as the set of materials to be recommended. When the quantity of materials in the second subset of materials to be recommended is less than the first threshold, the first subset of materials to be recommended is merged into the second subset of materials to be recommended, and materials whose quantity meets the first threshold are selected from the merged second subset of materials to be recommended as the set of materials to be recommended; wherein, the priority of the first subset of materials to be recommended is the next level below the priority of the second subset of materials to be recommended.

5. The method according to claim 1, characterized in that, Before determining the subset of materials to be recommended for each type of operation based on the recommendation strategy corresponding to different types of operation behaviors, the method further includes: Data cleaning is performed on the content data of the search behavior and / or the content data of the materials corresponding to the search behavior to filter out invalid content in the content data of the search behavior and / or the content data of the materials corresponding to the search behavior.

6. An information recommendation device, characterized in that, The device includes: The first acquisition module is used to acquire search behavior data of the target user, the search behavior data being used to indicate the search behavior of the target user; The first determining module is used to determine the set of materials corresponding to the search behavior based on the search behavior data; The second determining module is used to determine the type of operation behavior of the target user on the materials corresponding to the search behavior in the material set; The third determination module is used to determine the subset of materials to be recommended for each type of operation behavior according to the recommendation strategy corresponding to each type of operation behavior. The processing module is used to obtain a set of materials to be recommended based on the priority of the subset of materials to be recommended; The recommendation module is used to recommend materials to the target user based on the set of materials to be recommended; The target user's actions on the materials corresponding to the search behavior are used to indicate the target user's level of interest in the materials corresponding to the search behavior; the third determining module is used to: Based on the first operation behavior of the target user towards the materials corresponding to the search behavior, a first subset of materials to be recommended is determined according to the content data of the search behavior. Based on the second operation behavior of the target user towards the materials corresponding to the search behavior, a second subset of materials to be recommended is determined according to the content data of the materials corresponding to the search behavior; The materials corresponding to the search behavior are obtained from a preset material library based on the content data of the search behavior; the target user indicated by the second operation behavior has a higher degree of interest in the materials corresponding to the search behavior than the target user indicated by the first operation behavior. The third determining module is used for: Get the first set of the top N most popular content data with the highest search volume from the content data of the search behavior of the preset user, where N is an integer greater than 0; Obtain a second set of content data related to the search behavior of the target user; By taking the intersection of the first set and the second set, the popular content data of the target user's search behavior is obtained; Based on the popular content data, the first similarity between the popular content data of the target user's search behavior and the materials in the preset material library is calculated using a similarity algorithm model. The first subset of materials to be recommended is determined from the preset material library based on the first similarity.

7. An information recommendation device, characterized in that, The information recommendation device includes: Memory, used to store executable instructions; A processor, when executing executable instructions or computer programs stored in the memory, implements the method according to any one of claims 1 to 5.

8. A computer-readable storage medium storing executable instructions or a computer program, characterized in that, When the executable instructions are executed by the processor, they implement the method described in any one of claims 1 to 5.

Images