Content recommendation method and apparatus

Abstract

Embodiments of the present application provide a content recommendation method and device, comprising: obtaining first candidate content from a plurality of materials, the first candidate content corresponding to at least one content category label, and the first candidate content having click data; performing descending order sorting on the first candidate content corresponding to each content category label according to the click data, to obtain a candidate content ranking list corresponding to the content category label; for a plurality of first users publishing the first candidate content in the candidate content, selecting the first candidate content published by each first user and ranked in the top N according to the candidate content ranking list; determining a second user having an association relationship with at least one first user, and establishing a ranking list for the second user according to the click data of the first candidate content ranked in the top N of the at least one first user; and recommending a first target content for the second user according to the ranking list, the first target content being at least one of the first candidate content ranked in the top N of the at least one first user.

Description

Technical Field

[0001] This invention relates to the field of Internet technology, and in particular to a content recommendation method and apparatus. Background Technology

[0002] With the rapid development of the internet, more and more information is being published on social media platforms. Extracting and distributing materials that users might be interested in from this vast amount of information has become increasingly challenging. A common approach is to use models to filter large amounts of information, predict what information users might be interested in, and then distribute that information to re-engage users.

[0003] In some scenarios, a model is built based on user, material, and cross-feature characteristics to facilitate mixed-ranking and distribution. First, materials are recalled from a material pool containing millions of candidate materials. Different recall methods are used to retrieve materials of interest to users, which are then stored in the recall pool. During the offline model training phase, features are extracted and samples are generated from the materials in the material pool. These samples are then used for offline training. The materials recalled daily from various business segments are used for model calculation and ranking. Finally, the materials awaiting distribution for each business segment are mixed-ranked together to identify high-quality materials and distribute them to users for recall. However, this approach has limitations for low-frequency users. Due to the sparse behavior of low-frequency users and the limited number of candidate materials, ranking and mixed-ranking methods struggle to reach them effectively, making it difficult to recommend suitable materials and resulting in low user recall rates. Summary of the Invention

[0004] The purpose of this application is to provide a content recommendation method and apparatus to solve the problem of low user recall.

[0005] To solve the above-mentioned technical problems, the embodiments of this application are implemented as follows:

[0006] In a first aspect, embodiments of this application provide a content recommendation method, comprising: obtaining candidate content from multiple materials published by multiple first users, wherein each candidate content corresponds to at least one content category tag and has click data; for each first candidate content published by a first user, sorting the first candidate content corresponding to each content category tag in descending order according to the click data to obtain a candidate content ranking list corresponding to the content category tag; selecting the top N first candidate contents from the candidate content ranking list corresponding to each first user, where N is a natural number not less than 1; determining a second user associated with at least one first user, and establishing a ranking list for the second user based on the click data of the top N first candidate contents of the at least one first user; and recommending first target content to the second user based on the ranking list, wherein the first target content is at least one of the top N first candidate contents of the at least one first user.

[0007] Secondly, embodiments of this application provide a content recommendation device, comprising: an acquisition module, configured to acquire candidate content from multiple materials published by multiple first users, wherein the candidate content corresponds to at least one content category tag and has click data; a sorting module, configured to sort the first candidate content corresponding to each content category tag in descending order according to the click data for each first user's first candidate content, thereby obtaining a candidate content ranking list corresponding to the content category tag, wherein the candidate content includes the first candidate content; a selection module, configured to select the top N first candidate contents from the candidate content ranking list corresponding to each first user, wherein N is a natural number not less than 1; a determination module, configured to determine a second user associated with at least one first user, and establish a ranking list for the second user based on the click data of the top N first candidate contents of at least one first user; and a recommendation module, configured to recommend first target content to the second user according to the ranking list, wherein the first target content is at least one of the top N first candidate contents of at least one first user.

[0008] Thirdly, embodiments of this application provide an electronic device, including a processor, a communication interface, a memory, and a communication bus; wherein the processor, the communication interface, and the memory communicate with each other via the bus; the memory is used to store computer programs; and the processor is used to execute the programs stored in the memory to implement the steps as described in the first aspect.

[0009] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method steps as described in the first aspect.

[0010] Fifthly, embodiments of this application provide a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled, and the processor is used to run programs or instructions to implement the method steps as described in the first aspect.

[0011] As can be seen from the technical solutions provided in the embodiments of this application above, by obtaining candidate content from multiple materials published by multiple first users, each candidate content corresponds to at least one content category tag and has click data; for each first candidate content published by a first user, the first candidate content corresponding to each content category tag is sorted in descending order according to the click data to obtain a candidate content ranking list corresponding to the content category tag; the top N first candidate contents are selected from the candidate content ranking list corresponding to each first user, where N is a natural number not less than 1; a second user associated with at least one first user is determined, and a ranking list is established for the second user based on the click data of the top N first candidate contents of at least one first user; a first target content is recommended to the second user based on the ranking list, where the first target content is at least one of the top N first candidate contents of at least one first user. This solution can establish its own ranking list for each second user associated with a first user, thereby recommending target content to each user based on their own ranking list. This solution can cover low-frequency users, recommending suitable materials to low-frequency users based on their own ranking lists, thus improving user recall. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1 This is a schematic diagram of a first flowchart of a content recommendation method provided in an embodiment of this application;

[0014] Figure 2 A schematic diagram of the types of materials provided in the embodiments of this application;

[0015] Figure 3 A second flowchart illustrating the content recommendation method provided in this application embodiment;

[0016] Figure 4 A schematic diagram of the functional modules of the content recommendation device provided in the embodiments of this application;

[0017] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] The purpose of this application is to provide a content recommendation method and apparatus that can cover low-frequency users, thereby improving user recall.

[0019] To enable those skilled in the art to better understand the technical solutions of this invention, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this invention.

[0020] like Figure 1 As shown in the figure, this application embodiment provides a content recommendation method. The execution subject of this method can be a server, which can be an independent server or a server cluster composed of multiple servers. The content recommendation method specifically includes the following steps S101-S109:

[0021] In step S101, candidate content is obtained from multiple materials published by multiple first users.

[0022] Each candidate content corresponds to at least one content category tag and has click data.

[0023] Specifically, materials are obtained by mining and collecting features unique to certain scenarios, such as... Figure 2 As shown, the types of materials include trending materials, interest-based materials, and relationship-based materials. These materials each have corresponding content category tags, including trending tags for trending materials, interest tags for interest-based materials, and relationship tags for relationship-based materials. Trending materials refer to materials that quickly gain popularity, such as trending topics, current events, and materials that generate widespread discussion and receive numerous reposts, comments, likes, shares, and collections. Trending materials can be identified from a large pool of materials through content matching, click-through rate analysis, and rule matching. Interest-based materials refer to high-quality materials from specific content accounts, such as verified accounts (orange V, blue V, etc.). These are identified by identifying areas, vertical interests, and hobby circles that users may be interested in. This includes not only directly promoted materials but also materials manually reviewed by operations staff. Relationship-based materials refer to materials generated between users with mutual following relationships, such as mutual followers, friends on social media, classmates, relatives, and colleagues. These are identified through intimacy relationships; for example, if a user follows a blogger, all posts by that blogger will be added to that user's relationship material library. For these materials, there is corresponding click data, which includes, but is not limited to, click volume and click-through rate.

[0024] Candidate content is selected from the various materials mined above that meet the predetermined conditions. When selecting candidate content, it can be selected according to the click data of each material or according to the release time of each material. For example, materials whose release time does not exceed a preset time are selected as candidate content. The preset time can be 24 hours, etc.

[0025] In one possible implementation, obtaining candidate content from multiple materials published by multiple first users includes:

[0026] Retrieve multiple materials within a predetermined time period. Each material corresponds to at least one content category tag, which includes hot topic tags, interest tags, and relationship tags. Filter materials corresponding to interest tags that have click data exceeding a first threshold as candidate content, and / or filter materials corresponding to hot topic tags that have click data exceeding a second threshold as candidate content, and / or filter materials corresponding to relationship tags that have click data exceeding a third threshold as candidate content.

[0027] Specifically, when filtering materials corresponding to hot topic tags, the selection can be based on the number of clicks in the click data. Specifically, materials with clicks exceeding a second threshold are selected as candidate content. The second threshold can be determined based on actual conditions, and this embodiment does not limit its definition; for example, the second threshold could be set to 1000. When filtering materials corresponding to interest tags, the selection can be based on click-through rate or click volume. Specifically, the click-through rate or click volume of the materials can be sorted in descending order, and the top N materials are selected as candidate content, where N is a natural number not less than 1, such as N=200. When filtering materials corresponding to relationship tags, materials with click-through rates exceeding a third threshold in the click data can be selected as candidate content. The third threshold can be determined based on actual conditions, and this embodiment does not elaborate on its definition; for example, the third threshold could be 2%. Alternatively, when filtering materials corresponding to relationship tags, the selection can be based on the distribution volume, such as selecting materials with distribution volumes exceeding a preset value, where the preset value could be 500.

[0028] In one possible implementation, the material corresponding to the hot topic tag includes at least one of trending topics and trending current events; the material corresponding to the interest tag includes at least one of interest fields and hobby circles; and the material corresponding to the relationship tag includes material generated between users who have a following relationship.

[0029] In step S103, for each first candidate content published by the first user in the candidate content, the first candidate content corresponding to each content category tag is sorted in descending order according to the click data to obtain a candidate content ranking list corresponding to the content category tag.

[0030] Specifically, after filtering out hot topic candidate content corresponding to hot topic tags, interest candidate content corresponding to interest tags, and relationship candidate content corresponding to relationship tags, the first candidate content published by each first user is sorted in descending order according to click rate or click volume to obtain hot topic ranking list, interest ranking list, and relationship ranking list. The ranking list displays the identifier (such as the posting address) of the corresponding candidate content. The corresponding content is retrieved by the identifier and stored in Redis according to a specific format. It is then merged and deduplicated with the large number of relationship materials every day.

[0031] In step S105, the top N candidate contents are selected from the candidate content ranking list corresponding to each first user.

[0032] Where N is a natural number not less than 1.

[0033] Specifically, the first user can be the user who publishes the first candidate content, such as a blogger. Each blogger has its corresponding followers. For each first user, based on the aforementioned interest ranking list, hot topic ranking list, and relationship ranking list, the top N first candidate contents are selected from the interest ranking list, hot topic ranking list, and relationship ranking list for each first user. Each first user corresponds to the top N first candidate contents in the interest ranking list, hot topic ranking list, and relationship ranking list.

[0034] In step S107, a second user who is associated with at least one first user is identified, and a leaderboard is created for the second user based on the click data of the first candidate content ranked Nth by at least one first user.

[0035] Specifically, a second user refers to a user who has a following relationship and a closeness index with the first user, and the correspondence between the first and second users is stored. A second user may follow multiple first users, and each second user corresponds to one of the top N candidate content items for those first users. A ranking list is created for each second user based on click data for all the candidate content items corresponding to them. Specifically, the ranking list can be created based on the click-through rate or click volume of all candidate content items for each second user. When creating ranking lists for second users, a separate ranking list is created for each second user. When constructing rankings based on click-through rate and click volume, different ranking methods can be selected for different click volumes of materials, making it more suitable for ranking needs in different scenarios and improving ranking accuracy. It is important to note that when creating a leaderboard for each second user, the top N first-ranked candidate content from the first user's interest leaderboard, the top N first-ranked candidate content from the hot topic leaderboard, and the top N first-ranked candidate content from the relationship leaderboard are merged. A comprehensive leaderboard for each second user is then created based on the click-through rate or click-through volume of each of the top N first-ranked candidate content. This comprehensive leaderboard includes interest-related materials, hot topic materials, and relationship-related materials.

[0036] In one possible implementation, the click data includes click count and click-through rate. Building a leaderboard for a second user based on the click data of the top N first candidate content of at least one first user includes: determining the click count of the top N first candidate content of at least one first user; for the first candidate content of the top N first candidate content of at least one first user whose click count does not exceed a fourth threshold, building a leaderboard for the second user based on the click count of the first candidate content whose click count does not exceed the fourth threshold; and for the first candidate content of the top N first candidate content of at least one first user whose click count exceeds the fourth threshold, building a leaderboard for the second user based on the click-through rate of the first candidate content whose click count exceeds the fourth threshold.

[0037] Specifically, the click volume of at least one of the top N candidate items for a first user refers to: the click volume of the top N candidate items in the interest ranking list, the top N candidate items in the hot topic ranking list, and the top N candidate items in the relationship ranking list for the first user who is related to the second user. Among these candidate item click volumes, those with click volumes below a fourth threshold are ranked according to click volume, while those exceeding the fourth threshold are ranked according to click-through rate (CTR). The optimal candidate item is then selected from both the CTR and CTR ranking lists and recommended to the user.

[0038] In one possible implementation, before building a leaderboard for the second user based on the click-through rate of the first candidate content that exceeds a fourth threshold among the top N candidate content for at least one first user, the method further includes:

[0039] Determine the distribution volume for each first user and the click volume of the top N candidate content ranked by each first user; calculate the smoothing constant based on the total distribution volume determined by the distribution volume of each first user and the number of first users; calculate the click-through rate of each first candidate content based on the click volume of each first candidate content, the distribution volume of each first user, and the smoothing constant.

[0040] Specifically, the distribution volume refers to the total number of candidate content published by each first user, while the total distribution volume refers to the sum of the distribution volumes of all first users.

[0041] In one possible implementation, the click-through rate is calculated using the following formula:

[0042]

[0043] Where b is the smoothing constant, click i For the click count of the i-th first candidate content, push i Let be the amount of data distributed to the i-th first user.

[0044] in, n is the number of first users.

[0045] In step S109, the first target content is recommended to the second user based on the ranking list.

[0046] The first target content is at least one of the first candidate contents among the top N ranked by at least one first user.

[0047] Specifically, after determining the ranking list for the second user, the best candidate content from the overall ranking list is recommended to the user as the target content. Alternatively, one best candidate content is selected from both the ranking list based on click-through rate and the ranking list based on click volume as the target content to recommend to the user.

[0048] The technical solution disclosed in this application can establish a ranking list for each second user associated with the first user, thereby recommending target content to each user based on their own ranking list. This solution can cover low-frequency users and recommend suitable materials to them according to their own ranking list, thereby improving user recall.

[0049] In one possible implementation, such as Figure 3 As shown, after selecting the top N candidate contents from the candidate content leaderboard for each first user, the method further includes step S301:

[0050] In step S301, for each first user, a second target content is recommended for each first user based on the click data of the top N first candidate content in the candidate content ranking list.

[0051] Each second target content is at least one of the top N first candidate contents ranked for each first user.

[0052] Previously, for each first user, based on the aforementioned interest ranking list, hot topic ranking list, and relationship ranking list, the top N candidate contents were selected from each of these lists for each first user. Each first user had a corresponding top N candidate contents from the interest ranking list, hot topic ranking list, and relationship ranking list. These top N candidate contents were then ranked according to click volume or click-through rate to form a comprehensive ranking list for the first user. The comprehensive ranking list was sorted in descending order by click volume or click-through rate. The highest-ranked candidate contents from the comprehensive ranking list were selected as the second target content and recommended to the first user.

[0053] In this way, a leaderboard can be created for each user, and targeted content can be recommended to each user based on their own leaderboard. This solution can also cover low-frequency users, recommending suitable materials to them based on their own leaderboards, thereby improving user recall.

[0054] Corresponding to the content recommendation method provided in the above embodiments, based on the same technical concept, this application also provides a content recommendation device. Figure 4 This is a schematic diagram of the module composition of a content recommendation device provided in an embodiment of this application. This content recommendation device is used to execute the content recommendation method described in the above embodiments, such as... Figure 4As shown, the content recommendation device 400 includes: an acquisition module 401, used to acquire candidate content from multiple materials published by multiple first users, wherein the candidate content corresponds to at least one content category tag and the candidate content corresponds to click data; a sorting module 402, used to sort the first candidate content published by each first user in descending order according to the click data, thereby obtaining a candidate content ranking list corresponding to the content category tag, wherein the candidate content includes the first candidate content; a selection module 403, used to select the top N first candidate content from the candidate content ranking list corresponding to each first user, where N is a natural number not less than 1; a determination module 404, used to determine a second user associated with at least one first user, and to establish a ranking list for the second user based on the click data of the top N first candidate content of at least one first user; and a recommendation module 405, used to recommend first target content to the second user based on the ranking list, wherein the first target content is at least one of the top N first candidate content of at least one first user.

[0055] The technical solution disclosed in this application can establish a ranking list for each second user associated with the first user, thereby recommending target content to each user based on their own ranking list. This solution can cover low-frequency users and recommend suitable materials to them according to their own ranking list, thereby improving user recall rate.

[0056] In one possible implementation, the recommendation module 405 is further configured to recommend second target content for each first user based on the click data of the top N first candidate content in the candidate content ranking list, wherein each second target content is at least one of the top N first candidate content corresponding to each first user.

[0057] In one possible implementation, the acquisition module 401 is further configured to acquire multiple materials published by multiple first users within a predetermined time period, wherein the multiple materials correspond to at least one content category tag, and the content category tags include hot topic tags, interest tags, and relationship tags; sort the click data of the materials corresponding to the interest tags in descending order, filter the top N materials as candidate content, and / or filter the materials corresponding to the hot topic tags whose click data exceeds a second threshold as candidate content, and / or filter the materials corresponding to the relationship tags whose click data exceeds a third threshold as candidate content.

[0058] In one possible implementation, the material corresponding to the hot topic tag includes at least one of trending topics and trending current events; the material corresponding to the interest tag includes at least one of interest fields and hobby circles; and the material corresponding to the relationship tag includes material generated between users who have a following relationship.

[0059] In one possible implementation, the click data includes click count and click-through rate. The determining module 404 is further configured to determine the click count of the top N first candidate content of at least one first user; for the first candidate content of the top N first candidate content of at least one first user whose click count does not exceed a fourth threshold, a ranking list is established for the second user based on the click count of the first candidate content whose click count does not exceed the fourth threshold; for the first candidate content of the top N first candidate content of at least one first user whose click count exceeds the fourth threshold, a ranking list is established for the second user based on the click-through rate of the first candidate content whose click count exceeds the fourth threshold.

[0060] In one possible implementation, the determining module 404 is further configured to determine the distribution volume of each first user and the click volume of the top N first candidate content of each first user; calculate a smoothing constant based on the total distribution volume determined by the distribution volume of each first user and the number of first users; and calculate the click-through rate of each first candidate content based on the click volume of each first candidate content, the distribution volume of each first user, and the smoothing constant.

[0061] In one possible implementation, the click-through rate is calculated using the following formula:

[0062]

[0063] Where b is the smoothing constant, click i As a smoothing constant, click i For the click count of the i-th first candidate content, push i Let be the amount of data distributed to the i-th first user.

[0064] The content recommendation device provided in this application embodiment can implement the various processes in the embodiments corresponding to the above-described content recommendation method. To avoid repetition, it will not be described again here.

[0065] It should be noted that the content recommendation device provided in this application embodiment and the content recommendation method provided in this application embodiment are based on the same inventive concept and have the same technical effect. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned content recommendation method, and the repeated parts will not be described again.

[0066] Corresponding to the content recommendation method provided in the above embodiments, based on the same technical concept, this application also provides an electronic device for executing the above content recommendation method. Figure 5 To illustrate the structure of an electronic device according to various embodiments of the present invention, as shown in the schematic diagram... Figure 5As shown. Electronic devices can vary considerably due to differences in configuration or performance, and may include one or more processors 501 and memory 502. Memory 502 may store one or more application programs or data. Memory 502 may be temporary or persistent storage. The application programs stored in memory 502 may include one or more modules (not shown), each module may include a series of computer-executable instructions for the electronic device. Furthermore, processor 501 may be configured to communicate with memory 502 and execute the series of computer-executable instructions in memory 502 on the electronic device. The electronic device may also include one or more power supplies 503, one or more wired or wireless network interfaces 504, one or more input / output interfaces 505, and one or more keyboards 506.

[0067] In this embodiment, the electronic device includes a processor, a communication interface, a memory, and a communication bus; wherein, the processor, the communication interface, and the memory communicate with each other through the bus; the memory is used to store computer programs; the processor is used to execute the programs stored in the memory to implement the steps described in the above method embodiment.

[0068] It should be noted that the electronic device provided in this application embodiment and the content recommendation method provided in this application embodiment are based on the same inventive concept and have the same technical effect. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned content recommendation method, and the repeated parts will not be described again.

[0069] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps described in the above method embodiments.

[0070] It should be noted that the computer-readable storage medium provided in this application embodiment and the content recommendation method provided in this application embodiment are based on the same inventive concept and have the same technical effect. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned content recommendation method, and the repeated parts will not be described again.

[0071] In a specific embodiment, this application provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled, and the processor is used to run programs or instructions to implement the steps described in the above method embodiments.

[0072] It should be noted that the chip provided in this application embodiment and the content recommendation method provided in this application embodiment are based on the same inventive concept and have the same technical effect. Therefore, the specific implementation of this embodiment can refer to the implementation of the aforementioned content recommendation method, and the repeated parts will not be described again.

[0073] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, apparatus, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0074] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0075] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0076] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0077] In a typical configuration, an electronic device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0078] Memory may include non-persistent storage in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory is an example of computer-readable media.

[0079] Computer-readable media include both permanent and non-permanent, removable and non-removable media that can store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0080] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0081] Those skilled in the art will understand that embodiments of this application can be provided as methods, apparatus, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0082] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A content recommendation method characterized by, include: Candidate content is obtained from multiple materials published by multiple first users, and each candidate content corresponds to at least one content category tag and has click data. For each first candidate content published by the first user in the candidate content, the first candidate content corresponding to each content category tag is sorted in descending order according to the click data to obtain a candidate content ranking list corresponding to the content category tag; Select the top N candidate contents from the candidate content ranking list corresponding to each first user, where N is a natural number not less than 1; Identify a second user who is associated with at least one first user, and create a leaderboard for the second user based on the click data of the first candidate content ranked in the top N of the at least one first user; Based on the ranking list, a first target content is recommended to the second user, and the first target content is at least one of the first candidate contents ranked in the top N by the at least one first user; The click data includes click count and click-through rate. The step of building a leaderboard for the second user based on the click data of the top N first candidate content of the at least one first user includes: Determine the click count of the first candidate content in the top N of the ranking for at least one first user; For the first candidate content whose click volume does not exceed the fourth threshold among the top N first candidate content of the at least one first user, a ranking list is established for the second user based on the click volume of the first candidate content whose click volume does not exceed the fourth threshold; For the first candidate content whose click volume exceeds a fourth threshold among the top N first candidate content of the at least one first user, a ranking list is established for the second user based on the click rate of the first candidate content whose click volume exceeds the fourth threshold.

2. The content recommendation method according to claim 1, characterized by, After selecting the top N candidate contents from the candidate content ranking list corresponding to each first user, the method further includes: For each first user, a second target content is recommended for each first user based on the click data of the top N first candidate content in the candidate content ranking list, and each second target content is at least one of the top N first candidate content corresponding to each first user.

3. The content recommendation method of claim 1, wherein, The process of obtaining candidate content from multiple materials published by multiple first users includes: Get multiple materials published by multiple first users within a predetermined time period. The multiple materials correspond to at least one content category tag. The content category tag includes hot topic tags, interest tags, and relationship tags. The click data of the materials corresponding to the interest tags are sorted in descending order, and the top N materials are selected as candidate content, and / or the materials corresponding to the hotspot tags are selected as candidate content if their click data exceeds a second threshold, and / or the materials corresponding to the relationship tags are selected as candidate content if their click data exceeds a third threshold.

4. The content recommendation method according to claim 3, characterized by, The materials corresponding to the hot topic tags include at least one of trending topics and trending current events; the materials corresponding to the interest tags include at least one of interest fields and hobby circles; and the materials corresponding to the relationship tags include materials generated between users who have a following relationship.

5. The content recommendation method of claim 1, wherein, Before building a leaderboard for the second user, the method further includes, among the first candidate content in the top N ranked candidate content for at least one first user, the first candidate content whose click volume exceeds a fourth threshold, and based on the click-through rate of the first candidate content whose click volume exceeds the fourth threshold: Determine the distribution volume for each first user and the click volume of the first candidate content ranked in the top N for each first user; Calculate the smoothing constant based on the total amount of data sent to each of the first users and the number of the first users; The click-through rate of each first candidate content is calculated based on the number of clicks for each first candidate content, the number of downloads for each first user, and the smoothing constant.

6. The content recommendation method according to claim 5, characterized by, The click-through rate is calculated using the following formula: in, It is a smoothing constant. For the first The number of clicks on the first candidate content. For the first The amount of data distributed to the first user.

7. A content recommendation apparatus characterized by comprising: include: The acquisition module is used to acquire candidate content from multiple materials published by multiple first users. The candidate content corresponds to at least one content category tag and has click data. The click data includes the number of clicks and the click-through rate. The sorting module is used to sort the first candidate content corresponding to each content category tag in descending order according to the click data for each first candidate content published by the first user in the candidate content, so as to obtain a candidate content ranking list corresponding to the content category tag. The selection module is used to select the top N candidate contents from the candidate content ranking list corresponding to each first user, where N is a natural number not less than 1. The determination module is used to determine a second user who is associated with at least one first user, and to build a leaderboard for the second user based on the click data of the first candidate content ranked in the top N of the at least one first user. The recommendation module is used to recommend first target content to the second user based on the ranking list, wherein the first target content is at least one of the first candidate contents ranked in the top N by the at least one first user; The determining module is specifically used to determine the number of clicks on the first candidate content of the top N ranked items of the at least one first user; For the first candidate content whose click volume does not exceed the fourth threshold among the top N first candidate content of the at least one first user, a ranking list is established for the second user based on the click volume of the first candidate content whose click volume does not exceed the fourth threshold; For the first candidate content whose click volume exceeds a fourth threshold among the top N first candidate content of the at least one first user, a ranking list is established for the second user based on the click rate of the first candidate content whose click volume exceeds the fourth threshold.

8. An electronic device, comprising: include: processor; Memory used to store the processor's executable instructions; The processor is configured to execute the instructions to implement the content recommendation method as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The instructions in the storage medium, when executed by a processor of an electronic device, enable the electronic device to perform the content recommendation method of any one of claims 1 to 6.

Images