Content recommendation method, electronic device, storage medium and program product

Abstract

The invention discloses a content recommendation method, electronic equipment, a storage medium and a program product. The method comprises the steps of obtaining first account interaction data of a target account and content interaction data of multiple items of to-be-recommended content in response to a content recommendation request of the target account; determining a corresponding first candidate content set and a first recommended content set according to the first account interaction data, and determining target recommended content corresponding to the target account; acquiring account feature data of a target account, and generating a content recommendation reason corresponding to the target recommendation content according to the account feature data, content feature data corresponding to the target recommendation content and the content similarity between the target recommendation content and the second to-be-recommended content, the target recommendation content and the content recommendation reason corresponding to the target recommendation content are displayed in an associated manner, so that the accuracy of the recommendation content is improved, the recommendation information is enriched through the content recommendation reason, the recommendation content can be screened again conveniently, and the interaction experience of content recommendation is improved.

Description

Technical Field

[0001] The present invention relates to the field of computer application technology, and in particular to a content recommendation method, electronic equipment, storage medium and program product. Background Art

[0002] In the information age, although users can obtain a large amount of content through the Internet, it often takes a lot of time to filter out effective information that fits their needs. Therefore, how to quickly obtain effective information that fits their needs from a large amount of digital content has become one of the issues that more and more users are concerned about.

[0003] Currently, many applications have the function of content recommendation. However, the content recommendation method in the related art only feeds back the recommended content. The matching degree between the content and the requesting account is determined by the content recommendation method itself, and the matching degree between the recommended content and the user is unstable. Especially when the user cannot fully understand the recommended content, the user is often required to view the recommended content details to filter the information, which is time-consuming and labor-intensive, affecting the interactive experience. Summary of the invention

[0004] The present invention provides a content recommendation method, an electronic device, a storage medium and a program product to solve the problems in the related art of time-consuming and labor-intensive process of re-viewing recommended content by viewing details and poor interactive experience.

[0005] According to one aspect of the present invention, there is provided a content recommendation method, comprising:

[0006] In response to a content recommendation request from a target account, obtaining first account interaction data of the target account and content interaction data of a plurality of contents to be recommended; wherein the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account;

[0007] Determine a first candidate content set and a first recommended content set corresponding to the target account from a plurality of the to-be-recommended contents according to the first account interaction data, and determine a target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set; wherein the first candidate content set includes the first to-be-recommended content for the target account to select, and the first recommended content set includes the second to-be-recommended content matching the target account;

[0008] Acquire account characteristic data of the target account, generate a content recommendation reason corresponding to the target recommended content based on the account characteristic data and content characteristic data corresponding to the target recommended content, and content similarity between the target recommended content and the second content to be recommended, and associate the target recommended content with the content recommendation reason corresponding to the target recommended content for display.

[0009] According to another aspect of the present invention, there is provided a content recommendation device, comprising:

[0010] a content recommendation request module, configured to obtain, in response to a content recommendation request of a target account, first account interaction data of the target account and content interaction data of a plurality of contents to be recommended; wherein the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account;

[0011] a recommended content determination module, configured to determine, from the plurality of contents to be recommended, a first candidate content set and a first recommended content set corresponding to the target account according to the first account interaction data, and determine target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set; wherein the first candidate content set includes first contents to be recommended for selection by the target account, and the first recommended content set includes second contents to be recommended matching the target account;

[0012] The recommended content display module is used to obtain the account characteristic data of the target account, generate the content recommendation reason corresponding to the target recommended content according to the account characteristic data and the content characteristic data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, and associate the target recommended content with the content recommendation reason corresponding to the target recommended content for display.

[0013] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0014] at least one processor; and

[0015] a memory communicatively connected to the at least one processor; wherein,

[0016] The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the content recommendation method described in any embodiment of the present invention.

[0017] According to another aspect of the present invention, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the content recommendation method described in any embodiment of the present invention when executed.

[0018] According to another aspect of the present invention, an embodiment of the present disclosure further provides a computer program product, including a computer program, which implements the content recommendation method as described in any one of the embodiments of the present disclosure when executed by a processor.

[0019] The technical solution of the embodiment of the present invention is as follows: first, in response to a content recommendation request of a target account, first account interaction data of the target account and content interaction data of multiple contents to be recommended are obtained; since the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account, the recommendation needs of the user can be accurately located; then, a first candidate content set and a first recommended content set corresponding to the target account are determined from multiple contents to be recommended according to the first account interaction data; target recommended content corresponding to the target account is determined according to the first candidate content set and the first recommended content set; since the first candidate content set includes the first content to be recommended for selection by the target account, and the first recommended content set includes the second content to be recommended that matches the target account Recommended content can further optimize the recommended content structure and accurately meet user needs; finally, by obtaining the account feature data of the target account, the content recommendation reason corresponding to the target recommended content is generated according to the account feature data and the content feature data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, and the target recommended content and the content recommendation reason corresponding to the target recommended content are associated and displayed, thereby enhancing the transparency of the recommendation and improving the user's sense of trust. This solves the problems of the related technology of needing to view the details of the recommended content again, which leads to time-consuming and labor-intensive operations and poor interactive experience, and improves the accuracy and differentiation of the recommended content. By providing the target account with the content recommendation reason, the transparency of the recommendation method can be increased, and it is convenient for the target account to quickly understand the relevant information of the target recommended content, facilitate re-screening, and improve the interactive experience of the target account.

[0020] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0022] Figure 1 is a flowchart of a content recommendation method provided according to Embodiment 1 of the present invention;

[0023] Figure 2 is a flowchart of a content recommendation method provided according to Embodiment 2 of the present invention;

[0024] Figure 3 is a structural diagram of a content recommendation device provided according to Embodiment 3 of the present invention;

[0025] Figure 4 It is a structural schematic diagram of an electronic device for implementing the content recommendation method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0027] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0028] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0029] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only used for illustrative purposes and are not used to limit the scope of these messages or information.

[0030] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0031] For example, in response to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested to be performed will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, application, server, or storage medium that performs the operation of the technical solution of the present disclosure according to the prompt message.

[0032] As an optional but non-limiting implementation, in response to receiving an active request from the user, the prompt information may be sent to the user in the form of a pop-up window, in which the prompt information may be presented in text form. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0033] It is understandable that the above notification and the process of obtaining user authorization are merely illustrative and do not constitute a limitation on the implementation of the present disclosure. Other methods that meet the relevant laws and regulations may also be applied to the implementation of the present disclosure.

[0034] It is understandable that the data involved in this technical solution (including but not limited to the data itself, the acquisition or use of the data) shall comply with the requirements of relevant laws, regulations and relevant provisions.

[0035] Embodiment 1

[0036] Figure 1 A flowchart of a content recommendation method is provided for the first embodiment of the present invention. This embodiment is applicable to situations where differentiated content recommendations need to be made for target accounts. The method can be performed by a content recommendation device, which can be implemented in the form of hardware and / or software. Optionally, it can be implemented by an electronic device, which can be a mobile terminal, a PC, a server, etc.

[0037] like Figure 1 As shown, the method may specifically include:

[0038] S110. In response to a content recommendation request from a target account, obtain first account interaction data of the target account and content interaction data of multiple contents to be recommended; wherein the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account.

[0039] Among them, the target account can be understood as an account that receives content recommendation services. In this embodiment, the target account is the object of the differentiated recommendation service, and differentiated recommended content will be provided for it based on the difference information of the historical interaction data of the target account. The first account interaction data can be understood as the data of the target account and the content to be recommended (such as clicking, viewing details, applying or collecting, etc.), which is used to analyze the difference of the account interaction data and provide a basis for subsequent content recommendation, which may include but is not limited to at least one of the operations such as browsing history, liking and commenting. The content to be recommended can be understood as the content to be recommended to the target account, which may include but is not limited to at least one of articles, videos or messages. After screening, the content to be recommended becomes the candidate content that may be recommended to the target account. The content interaction data can be understood as the interaction data between the reference account) and the content to be recommended, and the reference account includes the account that has interacted with the content to be recommended within a preset time period. The reference account corresponding to the content to be recommended may include the target account, or may not include the target account. Through the content interaction data of the content to be recommended, the applicable account of the content to be recommended can be better understood, so as to further optimize the recommendation algorithm.

[0040] The content recommendation request is used to trigger the execution of the content recommendation process, and obtain the target recommended content corresponding to the target account and the reason for the content recommendation. In an optional implementation, in response to the target account opening the target interface operation, a content recommendation request corresponding to the target account is generated; or, after the interactive interface is triggered, multiple information setting items are displayed, and information setting operations are received by triggering the information setting items, and content recommendation requests are generated according to the information setting operations. Among them, the information setting items are used to set information associated with the expected recommended content. More specifically, the information setting items can be used to set one or more content filtering conditions corresponding to the target recommended content expected to be presented on the interactive interface.

[0041] S120. Determine a first candidate content set and a first recommended content set corresponding to the target account from multiple contents to be recommended according to the first account interaction data, and determine target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set; wherein the first candidate content set includes first content to be recommended for selection by the target account, and the first recommended content set includes second content to be recommended matching the target account.

[0042] Among them, the first candidate content set can be understood as a set of candidate content to be recommended that is screened out from a preset plurality of content to be recommended (all content to be recommended) based on the historical interaction data of the target account. The content of the first candidate content set can be considered as the content to be recommended that is preliminarily screened out from the preset plurality of content to be recommended and meets the interaction characteristics of the target account. The first recommended content set can be understood as a set of content to be recommended that has interactive operations (interacted) with the target account and matches the account interaction characteristics. Specifically, the second content to be recommended that matches the target account can be content to be recommended that has an interaction with the target account and the target interaction data meets the preset interaction conditions. The preset interaction conditions can include at least one of the interaction times restriction conditions, the interaction time restriction conditions, and the interaction mode restriction conditions. For example, the preset interaction conditions can include that the number of interactions with the content to be recommended in a preset interaction mode within a preset time period reaches a preset number of times, etc. The target recommended content can be understood as the content to be recommended that is recommended to the target account. The target recommended content is the final result of the recommendation process, and is the content with the account interaction characteristics determined after multiple screenings and optimizations.

[0043] It should be noted that, in the embodiment of the present invention, the “first”, “second” and “third” in the first content to be recommended, the second content to be recommended and the third content to be recommended are used to distinguish the content to be recommended in different sets, and are not limitations on specific content data such as the specific presentation data and presentation form of the recommended content.

[0044] In an optional implementation manner, a first candidate content set corresponding to the target account is determined from multiple contents to be recommended based on the first account interaction data, and the determination can be made in at least one of the following ways: obtaining the contents to be recommended that have interactive operations with the target account included in the first account interaction data, and classifying the contents to be recommended that have interactive operations with the target account to obtain interactive content categories, and determining the first candidate content set from a set of contents to be recommended corresponding to multiple preset content categories based on the interactive content categories; or, determining the number of interactions and interaction times of multiple preset contents to be recommended, determining multiple candidate contents to be recommended corresponding to the target account from multiple preset contents to be recommended based on the number of interactions and the interaction time, and constructing a first candidate content set based on the multiple candidate contents to be recommended, etc.

[0045] S130. Acquire account characteristic data of the target account, generate a content recommendation reason corresponding to the target recommended content according to the account characteristic data and content characteristic data corresponding to the target recommended content, and content similarity between the target recommended content and the second content to be recommended, and associate the target recommended content with the content recommendation reason corresponding to the target recommended content for display.

[0046] Among them, the account characteristic data can be understood as characteristic data used to distinguish different accounts. Exemplarily, the account characteristic data may include but are not limited to the subject and / or type of content. The account characteristic data helps to determine which content to be recommended is more suitable for recommendation to a specific account. The content similarity can be understood as an indicator to measure the degree of similarity between the target recommended content and the second content to be recommended. Combined with the content similarity between the target recommended content and the second content to be recommended, the reason for content recommendation to the target account can be more accurately determined. The content recommendation reason can be understood as explaining the reason why the target recommended content is recommended to the target account. Providing the content recommendation reason to the target account can increase the transparency of the recommendation method, and facilitate the target account to quickly understand the relevant information of the target recommended content, facilitate re-screening, and enhance the interactive experience of the target account.

[0047] On the basis of the above scheme, optionally, the generating of the content recommendation reason corresponding to the target recommended content according to the account characteristic data and the content characteristic data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, includes: inputting the account characteristic data, the content characteristic data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended into a recommendation reason generation model to obtain the content recommendation reason corresponding to the target recommended content; wherein, the recommendation reason generation model is obtained by training a deep learning model through content characteristic data of multiple second sample contents, the content similarities corresponding to the second sample contents, and sample account characteristics.

[0048] Among them, the recommendation reason generation model can be understood as a deep learning model for generating recommendation reasons. The recommendation reason generation model can generate recommendation reasons based on the account feature data and the content feature data corresponding to the target recommended content, the content similarity between the target recommended content and the second content to be recommended by learning a large amount of second sample content. The recommendation reason can be used to prompt the account of the specific reason why a certain recommended content is recommended to them, thereby increasing the transparency and credibility of the recommendation. The second sample content can be understood as sample content used to train the recommendation reason generation model. The second sample content and the rest of the data related to the recommendation reason can be input into the recommendation reason generation model to train the model so that the model can generate high-quality recommendation reasons. The sample account features can be understood as feature data of sample accounts used to train the recommendation reason generation model, which can be used to help the recommendation reason generation model understand the recommendation needs of different types of target accounts, thereby generating recommendation reasons that better meet the expectations of the target accounts.

[0049] The above scheme generates detailed content recommendation reasons by inputting account feature data, content feature data of target recommended content, content similarity and other information into the recommendation reason generation model, which not only improves the transparency and credibility of the recommendation, but also enhances the understanding and trust of the account.

[0050] The technical solution of the embodiment of the present invention is as follows: first, in response to a content recommendation request of a target account, first account interaction data of the target account and content interaction data of multiple contents to be recommended are obtained; since the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account, the account recommendation demand is accurately located; then, a first candidate content set and a first recommended content set corresponding to the target account are determined from multiple contents to be recommended according to the first account interaction data; target recommended content corresponding to the target account is determined according to the first candidate content set and the first recommended content set; since the first candidate content set includes the first content to be recommended for selection by the target account, the first recommended content set includes the first content to be recommended that matches the target account The second content to be recommended can further optimize the recommended content structure; finally, by obtaining the account characteristic data of the target account, the content recommendation reason corresponding to the target recommended content is generated according to the account characteristic data and the content characteristic data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, and the target recommended content and the content recommendation reason corresponding to the target recommended content are associated and displayed, thereby enhancing the transparency of the recommendation and improving the trust in the account, solving the problems of time-consuming and labor-intensive process of re-viewing the recommended content by viewing the details and the poor interactive experience in the related technology, improving the accuracy and differentiation of the recommended content, and providing the target account with the content recommendation reason to increase the transparency of the recommendation method, and facilitating the target account to quickly understand the relevant information of the target recommended content, facilitate re-screening, and improve the interactive experience of the target account.

[0051] Embodiment 2

[0052] Figure 2A flowchart of a content recommendation method provided for the second embodiment of the present invention. This embodiment is based on the above embodiment, and further refines the determination of the target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set. Optionally, the target recommended content corresponding to the target account is determined according to the first candidate content set and the first recommended content set, including: for each of the first content to be recommended, determining the second recommended content set corresponding to the first recommended content, constructing a third recommended content set based on the same content to be recommended in the first recommended content set and the second recommended content set, wherein the second recommended content set includes a preset number of items of the content to be recommended with the highest similarity; determining the predicted matching degree between the first content to be recommended and the target account according to the content similarity between the first content to be recommended and each third content to be recommended in the third recommended content set and the historical matching degree of the target account to each third content to be recommended; determining the target recommended content corresponding to the target account according to the predicted matching values ​​of multiple first content to be recommended. For specific implementation methods, please refer to the description of this embodiment. Among them, the technical features that are the same or similar to the above embodiments are not repeated here.

[0053] like Figure 2 As shown, the method may specifically include:

[0054] S210. In response to a content recommendation request from a target account, obtain first account interaction data of the target account and content interaction data of multiple contents to be recommended; wherein the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account.

[0055] S220. Determine a first candidate content set and a first recommended content set corresponding to the target account from multiple contents to be recommended according to the first account interaction data, determine a second recommended content set corresponding to the first recommended content for each of the first contents to be recommended, and construct a third recommended content set based on the identical contents to be recommended in the first recommended content set and the second recommended content set, wherein the first candidate content set includes the first contents to be recommended for selection by the target account, the first recommended content set includes the second contents to be recommended matching the target account, and the second recommended content set includes a preset number of contents to be recommended with the highest similarity.

[0056] The preset number can be understood as the total number of contents to be recommended in the second recommended content set. The preset number can be used to control the number of target recommended contents to ensure the quality and diversity of the target recommended contents.

[0057] In an embodiment of the present disclosure, optionally, determining a second recommended content set corresponding to the first recommended content may be by determining a recommended content category corresponding to the first recommended content, obtaining a preset number of contents to be recommended from a set of contents to be recommended corresponding to a plurality of preset content categories according to the recommended content category, and then constructing a second recommended content set based on the obtained preset number of contents to be recommended.

[0058] The third recommended content set can be understood as a set constructed based on the same to-be-recommended content in the first recommended content set and the second recommended content set. Specifically, the first recommended content set and the second recommended content set can be intersected to obtain the third recommended content set. It can be understood that the third recommended content set appears in both the first recommended content set and the second recommended content set, indicating that the content in the third recommended content set is highly matched with the content recommendation request of the target account, further improving the accuracy and reliability of the recommendation, and reducing the elements in the set, thereby improving the computing efficiency.

[0059] S230: Determine a predicted matching value between the first content to be recommended and the target account according to a content similarity between the first content to be recommended and each item of the third content to be recommended in the set of third recommended contents and a historical matching degree of the target account to each item of the third content to be recommended.

[0060] The historical matching degree can be understood as the historical matching value of the account for the recommended content, which can be determined by the account content recommendation request and historical operation data. The predicted matching value can be understood as the predicted value of measuring the matching degree between a first content to be recommended and the target account. The higher the predicted matching value, the higher the matching degree between the first content to be recommended and the target account.

[0061] On the basis of the above solution, specifically, the predicted matching value between the first content to be recommended and the target account is determined according to the content similarity between the first content to be recommended and each item of the third content to be recommended in the third recommended content set and the historical matching degree of the target account to each item of the third content to be recommended by the following formula:

[0062] R(u,i)=∑ j∈N(u)∩S(i,k) L ij ·W uj ,

[0063] Where u represents the target account, i represents the i-th first content to be recommended, j represents the j-th third content to be recommended, R(u,i) represents the predicted matching value of the target account u for the i-th first content to be recommended, N(u) represents the first recommended content set corresponding to the target account u, S(i,k) represents the second recommended content set corresponding to the target account u, W uj Represents the historical matching value between the target account u and the j-th third content to be recommended.

[0064] It can be understood that the historical matching value between the target account and the third content to be recommended can be determined based on the interaction data between the target account and the third content to be recommended in the first account interaction data.

[0065] By calculating the predicted matching value between the first content to be recommended and the target account, the content similarity and account historical operations can be comprehensively considered to generate more accurate and differentiated recommended content.

[0066] On the basis of the above scheme, optionally, before determining the predicted matching value of the first content to be recommended and the target account based on the content similarity between the first content to be recommended and each third content to be recommended in the third recommended content set and the historical matching degree of the target account to each third content to be recommended, it also includes: determining a reference account matching the first content to be recommended and a reference account matching the third content to be recommended based on the content interaction data, and determining the same reference account between the reference account matching the first content to be recommended and the reference account matching the third content to be recommended as a sample account; obtaining the second account interaction data of the sample account, and determining the interaction time between the sample account and the first content to be recommended and the interaction time between the sample account and the third content to be recommended based on the second account interaction data; determining the content similarity between the first content to be recommended and the third content to be recommended based on the interaction time between the sample account and the first content to be recommended, the interaction time between the sample account and the third content to be recommended, the total number of reference accounts matching the first content to be recommended, the total number of reference accounts matching the third content to be recommended, and the time influence factor.

[0067] Among them, the reference account can be understood as other accounts that have similar interactions or the same recommendation needs as the target account. More specifically, the reference account matching the first content to be recommended can be understood as an account that has interaction data with the first content to be recommended and meets the preset interaction conditions. Similarly, the reference account matching the third content to be recommended can be understood as an account that has interaction data with the third content to be recommended and meets the preset interaction conditions. The preset interaction conditions may include at least one of the interaction times restriction conditions, interaction time restriction conditions, and interaction mode restriction conditions. For example, the preset interaction conditions may include the number of interactions with the content to be recommended in a preset interaction mode within a preset time period reaching a preset number of times, etc. The sample account can be understood as the same reference account in the reference account matching the first content to be recommended and the reference account matching the third content to be recommended. The sample account is both an interaction account of the first content to be recommended and an interaction account of the third content to be recommended, and can be used to further calculate content similarity and predict matching values. The second account interaction data can be understood as a historical record of the sample account's interaction with the content on the interactive platform, including but not limited to at least one of browsing, clicking, resource transfer or commenting operations, and is used to analyze the sample account's interaction mode with the first content to be recommended and the third content to be recommended, and further optimize the recommendation algorithm. The interaction time can be understood as the time point when the sample account interacts with the first content to be recommended or the third content to be recommended, which is used to reflect the account's attention to the recommended content and can be used to calculate content similarity and predicted matching values. A more recent interaction time may indicate that the account currently pays more attention to the recommended content. The time impact factor can be understood as a parameter used to adjust the impact of the interaction time on content similarity and predicted matching values. It can be understood that as time continues to increase, the result of the time impact factor tends to 0, that is, the account gradually loses its enthusiasm for the recommended content, and the probability of being recommended also decreases.

[0068] On the basis of the above scheme, specifically, the content similarity between the first content to be recommended and the third content to be recommended can be determined by the following formula according to the interaction time between the sample account and the first content to be recommended, the interaction time between the sample account and the third content to be recommended, the total number of reference accounts matching the first content to be recommended, the total number of reference accounts matching the third content to be recommended, and the time impact factor:

[0069]

[0070] Among them, L ij represents the content similarity between the first recommended content of the ith item and the third recommended content of the jth item, u represents the sample account, B(i) represents the total number of reference accounts matching the first recommended content of the ith item, B(j) represents the total number of reference accounts matching the third recommended content of the jth item, β represents the time impact factor, tui represents the interaction time between the reference account u and the first recommended content of the i-th item, t uj Represents the interaction time between the reference account u and the j-th third content to be recommended.

[0071] An optional implementation method, since the business scenario of product recommendation is a situation where the number of accounts is large and the number of product categories is relatively small, a product-based collaborative filtering algorithm is used for recall. First, the similarity between different products (such as product i and product j, B(i) and B(j) are the number of accounts with content recommendation requirements) is calculated:

[0072]

[0073] The above scheme, by comprehensively considering the interaction time of sample accounts, the number of reference accounts and the time influence factor, can more accurately calculate the content similarity between the first content to be recommended and the third content to be recommended, fully considering the relevance and difference of different contents to be recommended, and also considering the importance of interaction time in the recommendation process and the degree of influence on the recommendation results. The time influence factor reduces the impact of premature interaction records, so that the recommendation process pays more attention to the account's most recent interaction characteristics and is more in line with the account interaction characteristics of the target account, thereby improving the matching degree between the target recommended content and the target account.

[0074] S240: Determine target recommended content corresponding to the target account according to the predicted matching values ​​of the plurality of first content to be recommended.

[0075] Specifically, multiple fourth contents to be recommended are determined based on the predicted matching values ​​of multiple first contents to be recommended, and the multiple fourth contents to be recommended are reordered through the content sorting model to obtain a recommendation order corresponding to each of the fourth contents to be recommended, and the target recommended content corresponding to the target account is determined based on the recommendation order corresponding to the fourth contents to be recommended.

[0076] The fourth content to be recommended can be understood as content selected based on the predicted matching values ​​of the first content to be recommended. The fourth content to be recommended is the content that is considered to be most likely to meet the recommendation needs of the target account after preliminary screening, and needs to be sorted by the content sorting model to determine the final recommendation order.

[0077] Exemplarily, multiple fourth contents to be recommended are determined based on the predicted matching values ​​of multiple first contents to be recommended. This may be by determining a preset number of the first contents to be recommended with the highest predicted matching values ​​as the fourth contents to be recommended, or by determining a preset number of the first contents to be recommended with predicted matching values ​​higher than a preset matching threshold as the fourth contents to be recommended, etc.

[0078] In an embodiment of the present invention, the content ranking model can be understood as a model for ranking recommended content. The content ranking model can re-rank the input content to be recommended according to the characteristics of the first sample content and the operation mode of the account by learning multiple sample data, so as to optimize the recommendation order and improve the satisfaction of the target account with the target recommended content. The content ranking model can be obtained by training the extreme gradient boosting model (ExtremeGradient Boosting Model, XGBoost) through multiple first sample contents. The extreme gradient boosting model can be understood as a machine learning model that is trained by the gradient boosting algorithm and is suitable for classification and regression tasks. The extreme gradient boosting model constructs a strong learner by combining multiple weak learners (usually decision trees), which can process large-scale data sets and provide high-precision predictions. In this embodiment, the extreme gradient boosting model is often used to train the content ranking model. The first sample content can be understood as the sample content used to train the content ranking model, so that it can learn which content characteristics and account operation modes have the greatest impact on the recommendation order, so that it can accurately sort the input content to be recommended in the application stage.

[0079] In an optional implementation, the extreme gradient boosting model can be used to reorder multiple fourth contents to be recommended, and the reordering prediction results of multiple fourth contents to be recommended can be output. Furthermore, the protocol determines the fourth contents to be recommended that are ranked at the front (with the highest priority in the recommended order) as the target recommended contents. The adoption of this technical solution can further improve the recommendation accuracy of the contents to be recommended.

[0080] S250, obtaining the account characteristic data of the target account, generating a content recommendation reason corresponding to the target recommended content according to the account characteristic data and the content characteristic data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, and displaying the target recommended content and the content recommendation reason corresponding to the target recommended content in association with each other.

[0081] The technical solution of the embodiment of the present invention determines the first candidate content set and the first recommended content set through the first account interaction data and the content interaction data, so as to ensure the preliminary screening quality of the recommended content. Then, the predicted matching value is calculated by using the content similarity and the historical matching degree, and the fourth content to be recommended is further screened out. The fourth content to be recommended is re-sorted through the content sorting model, and the recommendation order is optimized to ensure the best final presentation effect of the recommended content. This not only improves the accuracy and differentiation of the recommended content, but also optimizes the recommendation process, and improves the account experience and the interaction rate of the platform.

[0082] Embodiment 3

[0083] Figure 3 This is a schematic diagram of the structure of a content recommendation device provided by Embodiment 3 of the present invention. Figure 3 As shown, the device includes: a content recommendation request module 310, a recommended content determination module 320, and a recommended content display module 330. Among them, the content recommendation request module 310 is used to respond to the content recommendation request of the target account and obtain the first account interaction data of the target account and the content interaction data of multiple contents to be recommended; wherein the first account interaction data at least includes the interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes the interaction data with at least one reference account; the recommended content determination module 320 is used to determine the first candidate content set and the first recommended content set corresponding to the target account from the multiple contents to be recommended according to the first account interaction data, and determine the target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set; wherein the first candidate content set includes the first content to be recommended for selection by the target account, and the first recommended content set includes the second content to be recommended matching the target account; the recommended content display module 330 is used to obtain the account feature data of the target account, generate the content recommendation reason corresponding to the target recommended content according to the account feature data and the content feature data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, and associate the target recommended content with the content recommendation reason corresponding to the target recommended content.

[0084] The technical solution of the embodiment of the present invention is as follows: first, in response to the content recommendation request of the target account, the content recommendation request module 310 obtains the first account interaction data of the target account and the content interaction data of multiple contents to be recommended. Since the first account interaction data at least includes the interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes the interaction data with at least one reference account, the account recommendation demand is accurately located; then, the recommended content determination module 320 determines the first candidate content set and the first recommended content set corresponding to the target account from the multiple contents to be recommended according to the first account interaction data, and determines the target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set. Since the first candidate content set includes the first content to be recommended for the target account to select, the first recommended content set includes the content that matches the target account. The second content to be recommended can further optimize the recommended content structure to accurately meet the account needs; finally, the account feature data of the target account is obtained through the recommended content determination module 330, and the content recommendation reason corresponding to the target recommended content is generated according to the account feature data and the content feature data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, and the target recommended content and the content recommendation reason corresponding to the target recommended content are associated and displayed, so as to enhance the transparency of the recommendation and improve the trust in the account, solve the problems of time-consuming and labor-intensive process and poor interactive experience caused by the need to view the details of the recommended content in the related technology, improve the accuracy and differentiation of the recommended content, and provide the target account with the content recommendation reason to increase the transparency of the recommendation method, and facilitate the target account to quickly understand the relevant information of the target recommended content, facilitate re-screening, and enhance the interactive experience of the target account.

[0085] On the basis of the above-mentioned optional technical solutions, optionally, the recommended content determination module includes: a third recommended content set construction submodule, a predicted matching value determination submodule and a target recommended content determination submodule. The third recommended content set construction submodule is used to determine the second recommended content set corresponding to the first recommended content for each of the first content to be recommended, and to construct a third recommended content set based on the same content to be recommended in the first recommended content set and the second recommended content set, wherein the second recommended content set includes a preset number of content to be recommended with the highest similarity; the predicted matching value determination submodule is used to determine the predicted matching value of the first content to be recommended and the target account according to the content similarity between the first content to be recommended and each third content to be recommended in the third recommended content set and the historical matching degree of the target account to each third content to be recommended; the target recommended content determination submodule is used to determine the target recommended content corresponding to the target account according to the predicted matching values ​​of multiple first content to be recommended.

[0086] Based on the above optional technical solutions, optionally, the content recommendation device further includes: a sample account determination module, an interaction time determination module and a content similarity determination module. Among them, the sample account determination module is used to determine the reference account matching the first content to be recommended and the reference account matching the third content to be recommended according to the content interaction data, and determine the same reference account between the reference account matching the first content to be recommended and the reference account matching the third content to be recommended as the sample account, before determining the predicted matching value between the first content to be recommended and the target account according to the content similarity between the first content to be recommended and each third content to be recommended in the third recommended content set and the historical matching degree of the target account to each third content to be recommended; the interaction time determination module is used to obtain the second account interaction data of the sample account, and determine the interaction time between the sample account and the first content to be recommended and the interaction time between the sample account and the third content to be recommended according to the second account interaction data; the content similarity determination module is used to determine the content similarity between the first content to be recommended and the third content to be recommended according to the interaction time between the sample account and the first content to be recommended, the interaction time between the sample account and the third content to be recommended, the total number of reference accounts matching the first content to be recommended, the total number of reference accounts matching the third content to be recommended, and the time influence factor.

[0087] Based on the above optional technical solutions, optionally, the content similarity determination module is specifically used to: determine the content similarity between the first content to be recommended and the third content to be recommended according to the interaction time between the sample account and the first content to be recommended, the interaction time between the sample account and the third content to be recommended, the total number of reference accounts matching the first content to be recommended, the total number of reference accounts matching the third content to be recommended, and the time influence factor through the following formula:

[0088]

[0089] Among them, L ij represents the content similarity between the first recommended content of the ith item and the third recommended content of the jth item, u represents the sample account, B(i) represents the total number of reference accounts matching the first recommended content of the ith item, B(j) represents the total number of reference accounts matching the third recommended content of the jth item, β represents the time impact factor, t ui represents the interaction time between the reference account u and the first recommended content of the i-th item, t uj Represents the interaction time between the reference account u and the j-th third content to be recommended.

[0090] Based on the above optional technical solutions, optionally, the target recommended content determination submodule includes: a target recommended content determination unit. The target recommended content determination unit is used to determine a plurality of fourth contents to be recommended based on the predicted matching values ​​of the plurality of first contents to be recommended, reorder the plurality of fourth contents to be recommended through a content ranking model to obtain a recommendation order corresponding to each of the fourth contents to be recommended, and determine the target recommended content corresponding to the target account according to the recommendation order corresponding to the fourth contents to be recommended, wherein the content ranking model is obtained by training the extreme gradient boosting model through a plurality of first sample contents.

[0091] Based on the above optional technical solutions, optionally, the predicted matching value determination submodule is specifically used to: determine the predicted matching value between the first content to be recommended and the target account according to the content similarity between the first content to be recommended and each third content to be recommended in the third recommended content set and the historical matching degree of the target account to each third content to be recommended by the following formula:

[0092] R(u,i)=∑ j∈N(u)∩S(i,k) L ij ·W uj ,

[0093] Where u represents the target account, i represents the i-th first content to be recommended, j represents the j-th third content to be recommended, R(u,i) represents the predicted matching value of the target account u for the i-th first content to be recommended, N(u) represents the first recommended content set corresponding to the target account u, S(i,k) represents the second recommended content set corresponding to the target account u, W uj Represents the historical matching value between the target account u and the j-th third content to be recommended.

[0094] Based on the above optional technical solutions, optionally, the recommended content display module includes: a content recommendation reason determination submodule. The content recommendation reason determination submodule is used to input the account feature data, the content feature data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended into a recommendation reason generation model to obtain the content recommendation reason corresponding to the target recommended content; wherein the recommendation reason generation model is obtained by training a deep learning model through the content feature data of multiple second sample contents, the content similarity corresponding to the second sample contents, and the sample account features.

[0095] The content recommendation device provided in the embodiment of the present invention can execute the content recommendation method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0096] Embodiment 4

[0097] Figure 4 A schematic diagram of the structure of an electronic device 10 that can be used to implement an embodiment of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.

[0098] like Figure 4As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0099] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0100] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 performs the various methods and processes described above, such as the content recommendation method.

[0101] In some embodiments, the content recommendation method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the content recommendation method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the content recommendation method in any other appropriate manner (e.g., by means of firmware).

[0102] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0103] Computer programs for implementing the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0104] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0105] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).

[0106] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.

[0107] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.

[0108] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.

[0109] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A content recommendation method, characterized in that: include: In response to a content recommendation request from a target account, obtaining first account interaction data of the target account and content interaction data of a plurality of contents to be recommended; wherein the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account; Determine a first candidate content set and a first recommended content set corresponding to the target account from a plurality of the to-be-recommended contents according to the first account interaction data, and determine a target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set; wherein the first candidate content set includes the first to-be-recommended content for the target account to select, and the first recommended content set includes the second to-be-recommended content matching the target account; Acquire account characteristic data of the target account, generate a content recommendation reason corresponding to the target recommended content based on the account characteristic data and content characteristic data corresponding to the target recommended content, and content similarity between the target recommended content and the second content to be recommended, and associate the target recommended content with the content recommendation reason corresponding to the target recommended content for display.

2. The method according to claim 1, characterized in that The determining, according to the first candidate content set and the first recommended content set, target recommended content corresponding to the target account includes: For each of the first contents to be recommended, determine a second recommended content set corresponding to the first recommended content, and construct a third recommended content set based on the identical contents to be recommended in the first recommended content set and the second recommended content set, wherein the second recommended content set includes a preset number of contents to be recommended with the highest similarity; Determining a predicted matching value between the first content to be recommended and the target account according to the content similarity between the first content to be recommended and each item of the third content to be recommended in the third recommended content set and the historical matching degree of the target account to each item of the third content to be recommended; The target recommended content corresponding to the target account is determined according to the predicted matching values ​​of the plurality of first contents to be recommended.

3. The method according to claim 2, characterized in that Before determining the predicted matching value between the first content to be recommended and the target account according to the content similarity between the first content to be recommended and each item of the third content to be recommended in the set of third recommended contents and the historical matching degree of the target account to each item of the third content to be recommended, the method further includes: Determine, according to the content interaction data, a reference account matching the first content to be recommended and a reference account matching the third content to be recommended, and determine the same reference account between the reference account matching the first content to be recommended and the reference account matching the third content to be recommended as a sample account; Acquire second account interaction data of the sample account, and determine the interaction time between the sample account and the first content to be recommended and the interaction time between the sample account and the third content to be recommended according to the second account interaction data; The content similarity between the first content to be recommended and the third content to be recommended is determined based on the interaction time between the sample account and the first content to be recommended, the interaction time between the sample account and the third content to be recommended, the total number of reference accounts matching the first content to be recommended, the total number of reference accounts matching the third content to be recommended, and a time influence factor.

4. The method according to claim 3, characterized in that The content similarity between the first content to be recommended and the third content to be recommended is determined by the following formula according to the interaction time between the sample account and the first content to be recommended, the interaction time between the sample account and the third content to be recommended, the total number of reference accounts matching the first content to be recommended, the total number of reference accounts matching the third content to be recommended, and the time influence factor: Among them, L ij represents the content similarity between the first recommended content of the ith item and the third recommended content of the jth item, u represents the sample account, B(i) represents the total number of reference accounts matching the first recommended content of the ith item, B(j) represents the total number of reference accounts matching the third recommended content of the jth item, β represents the time impact factor, t ui represents the interaction time between the reference account u and the first recommended content of the i-th item, t uj Represents the interaction time between the reference account u and the j-th third content to be recommended.

5. The method according to claim 2, characterized in that: The determining the target recommended content corresponding to the target account according to the predicted matching values ​​of the plurality of first contents to be recommended includes: A plurality of fourth contents to be recommended are determined based on predicted matching values ​​of the plurality of first contents to be recommended, and the plurality of fourth contents to be recommended are reordered through a content ranking model to obtain a recommendation order corresponding to each of the fourth contents to be recommended, and a target recommended content corresponding to the target account is determined based on the recommendation order corresponding to the fourth contents to be recommended, wherein the content ranking model is obtained by training an extreme gradient boosting model through the plurality of first sample contents.

6. The method according to claim 2, characterized in that The predicted matching value between the first content to be recommended and the target account is determined by the following formula according to the content similarity between the first content to be recommended and each item of the third content to be recommended in the third recommended content set and the historical matching degree of the target account to each item of the third content to be recommended: R(u,i)=∑ j∈N(u)∩S(i,k) L ij ·W uj , Where u represents the target account, i represents the i-th first content to be recommended, j represents the j-th third content to be recommended, R(u,i) represents the predicted matching value of the target account u for the i-th first content to be recommended, N(u) represents the first recommended content set corresponding to the target account u, S(i,k) represents the second recommended content set corresponding to the target account u, W uj Represents the historical matching value between the target account u and the j-th third content to be recommended.

7. The method according to claim 1, characterized in that The generating a content recommendation reason corresponding to the target recommended content according to the account feature data, the content feature data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended includes: The account feature data, the content feature data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended are input into a recommendation reason generation model to obtain the content recommendation reason corresponding to the target recommended content; wherein the recommendation reason generation model is obtained by training a deep learning model through content feature data of multiple second sample contents, content similarities corresponding to the second sample contents, and sample account features.

8. A content recommendation device, characterized in that: include: a content recommendation request module, configured to obtain, in response to a content recommendation request of a target account, first account interaction data of the target account and content interaction data of a plurality of contents to be recommended; wherein the first account interaction data at least includes interaction data between the target account and at least one of the contents to be recommended, and the content interaction data at least includes interaction data with at least one reference account; a recommended content determination module, configured to determine, from the plurality of contents to be recommended, a first candidate content set and a first recommended content set corresponding to the target account according to the first account interaction data, and determine target recommended content corresponding to the target account according to the first candidate content set and the first recommended content set; wherein the first candidate content set includes first contents to be recommended for selection by the target account, and the first recommended content set includes second contents to be recommended matching the target account; The recommended content display module is used to obtain the account characteristic data of the target account, generate the content recommendation reason corresponding to the target recommended content according to the account characteristic data and the content characteristic data corresponding to the target recommended content, and the content similarity between the target recommended content and the second content to be recommended, and associate the target recommended content with the content recommendation reason corresponding to the target recommended content for display.

9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the content recommendation method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the content recommendation method according to any one of claims 1 to 7 when executed.

11. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the computer program implements the content recommendation method according to any one of claims 1 to 7.

Images