CN116932866A

CN116932866A - Content recommendation method, content recommendation device, storage medium, electronic equipment and product

Info

Publication number: CN116932866A
Application number: CN202210334168.0A
Authority: CN
Inventors: 刘刚
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2022-03-30
Filing date: 2022-03-30
Publication date: 2023-10-24
Anticipated expiration: 2042-03-30
Also published as: CN116932866B

Abstract

The application discloses a content recommendation method, a device, a storage medium, electronic equipment and a product, which relate to the technical field of artificial intelligence, and can be applied to the technical fields of block chains, map Internet of vehicles and the like, wherein the method is applied to a terminal and comprises the following steps: detecting content information corresponding to local browsing content, and detecting interaction behavior information corresponding to local content browsing interaction behavior; acquiring candidate recommended contents issued by a server, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request; analyzing and processing based on the content information and the interactive behavior information to obtain target browsing information; and carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation. The content recommendation method and device can improve content recommendation effects.

Description

Content recommendation method, content recommendation device, storage medium, electronic equipment and product

Technical Field

The application relates to the technical field of artificial intelligence, in particular to a content recommendation method, a content recommendation device, a storage medium, electronic equipment and a product.

Background

Content recommendation is a task of recommending content such as videos, articles, cards and the like, for example, a recommendation system recommends different types of content to a terminal for display in the process of browsing the content by a user so as to enable the user to browse.

At present, in a content recommendation scheme in a recommendation system, a terminal generally sends a request to a server (such as a cloud server), and the server decides a target recommendation content according to specific information and sends the target recommendation content to the terminal for recommendation.

In the current scheme, due to reasons of network bandwidth, delay and the like, specific information for decision making in a server is substantially earlier than the time for recommending target content in a terminal, that is, the server cannot effectively sense the real-time browsing condition in the terminal to decide the target recommended content, and the target recommended content is usually deviated from the real-time real demand of a browsing object, so that the content recommending effect is poor.

Disclosure of Invention

The embodiment of the application provides a content recommendation method and a related device, which can effectively improve the content recommendation effect.

In order to solve the technical problems, the embodiment of the application provides the following technical scheme:

according to one embodiment of the present application, a content recommendation method is applied to a terminal, and the method includes: detecting content information corresponding to local browsing content, and detecting interaction behavior information corresponding to local content browsing interaction behavior; acquiring candidate recommended contents issued by a server, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request; analyzing and processing based on the content information and the interactive behavior information to obtain target browsing information; and carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

According to an embodiment of the present application, a content recommendation apparatus, applied to a terminal, includes: the detection module is used for detecting content information corresponding to local browsing content and detecting interaction behavior information corresponding to local content browsing interaction behavior; the acquisition module is used for acquiring candidate recommended contents issued by the server, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request; the analysis module is used for carrying out analysis processing based on the content information and the interaction behavior information to obtain target browsing information; and the adjustment module is used for carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

In some embodiments of the present application, the content browsing interaction behavior includes a content exposure behavior and a browsing interaction behavior of a browsing object for browsing content; the detection module comprises: the exposure behavior information detection unit is used for detecting the local content exposure behavior to obtain content exposure behavior information; the browsing interaction behavior detection unit is used for detecting the local browsing interaction behavior to obtain browsing interaction behavior information; and the information determining unit is used for taking the content exposure behavior information and the browsing interaction behavior information as the interaction behavior information.

In some embodiments of the application, the content exposure behavior includes a positive feedback content exposure behavior and a negative feedback content exposure behavior; the exposure behavior information detection unit is used for: detecting the local positive feedback content exposure behavior to obtain positive feedback interaction information; detecting the local exposure behavior of the negative feedback content to obtain negative feedback interaction information; and taking the positive feedback interaction information and the negative feedback interaction information as the content exposure behavior characteristic information.

In some embodiments of the present application, the browsing interaction behavior includes a true browsing interaction behavior and a pseudo browsing interaction behavior; the browse interaction behavior detection unit is used for: detecting the local real browsing interaction behavior to obtain real browsing interaction behavior information; detecting the local pseudo-browsing interaction behavior to obtain pseudo-browsing interaction behavior information; and taking the real browsing interaction behavior information and the pseudo browsing interaction behavior information as the browsing interaction behavior information.

In some embodiments of the application, the analysis module comprises: the first extraction unit is used for carrying out feature extraction processing on the content information to obtain content features; the second extraction unit is used for carrying out feature extraction processing on the interaction behavior information to obtain interaction behavior features; and the analysis processing unit is used for carrying out analysis processing based on the interactive behavior characteristics and the content characteristics to obtain the target browsing information.

In some embodiments of the present application, the interactive behavior information includes content exposure behavior information and browsing interactive behavior information, and the interactive behavior features include exposure behavior features and browsing behavior features; the second extraction unit is used for: performing feature extraction processing on the content exposure behavior information to obtain the exposure behavior feature; and carrying out feature extraction processing on the browsing interaction behavior information to obtain the browsing behavior features.

In some embodiments of the application, the analysis processing unit comprises: the selecting subunit is used for selecting the content to be recommended from the candidate recommended contents and extracting the characteristics of the content to be recommended corresponding to the content to be recommended; the fusion subunit is used for carrying out fusion processing on the interactive behavior characteristics and the content characteristics to obtain fusion characteristics; the mutual fusion coding subunit is used for carrying out mutual fusion coding processing on the interactive behavior characteristics, the content characteristics to be recommended and the fusion characteristics to obtain characteristics to be identified; and the prediction subunit is used for performing prediction processing based on the features to be identified to obtain the target browsing information.

In some embodiments of the application, the intersolubility coding subunit is configured to: and performing self-attention operation processing by taking the content features to be recommended as query features, the interaction behavior features as key features and the fusion features as value features to obtain the features to be identified.

In some embodiments of the application, the adjustment module comprises: the strategy determining unit is used for determining a dynamic recommendation strategy of the recommended content in the candidate recommended content according to the target browsing information; and the adjusting unit is used for adjusting the recommended content in the candidate recommended content according to the dynamic recommended strategy to obtain the target recommended content.

In some embodiments of the application, the adjusting unit is configured to: determining recommended contents to be browsed from recommended contents included in the candidate recommended contents; and adjusting the recommended content to be browsed according to the dynamic recommendation strategy to obtain the target recommended content.

In some embodiments of the application, the apparatus further comprises: and the display module is used for recommending the recommended content in the candidate recommended content according to a preset recommendation strategy corresponding to the candidate recommended content.

According to one embodiment of the present application, a content recommendation method is applied to a server, and the method includes: receiving a content browsing request sent by a terminal; responding to the content browsing request to perform analysis processing to obtain candidate recommended content; and sending the candidate recommended content to the terminal so that the terminal can adjust the candidate recommended content according to target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation, and the target browsing information is obtained by analyzing and processing the terminal based on the interaction behavior information corresponding to the detected content browsing interaction behavior and the content information corresponding to the browsing content.

According to an embodiment of the present application, a content recommendation apparatus, applied to a server, includes: the receiving module is used for receiving a content browsing request sent by the terminal; the decision module is used for responding to the content browsing request to perform analysis processing so as to obtain candidate recommended content; the sending module is used for sending the candidate recommended content to the terminal so that the terminal can adjust the candidate recommended content according to target browsing information to obtain target recommended content, the target recommended content is used for recommending content, and the target browsing information is obtained by analyzing and processing the terminal based on the interaction behavior information corresponding to the detected content browsing interaction behavior and the content information corresponding to the browsing content.

According to another embodiment of the application, a computer readable storage medium has stored thereon a computer program which, when executed by a processor of a computer, causes the computer to perform the method according to the embodiment of the application.

According to another embodiment of the present application, an electronic device includes: a memory storing a computer program; and the processor reads the computer program stored in the memory to execute the method according to the embodiment of the application.

According to another embodiment of the application, a computer program product or computer program includes computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions to cause the computer device to perform the methods provided in the various alternative implementations described in the embodiments of the present application.

In the embodiment of the application, the terminal can: detecting content information corresponding to local browsing content, and detecting interaction behavior information corresponding to local content browsing interaction behavior; acquiring candidate recommended contents issued by a server, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request; analyzing and processing based on the content information and the interactive behavior information to obtain target browsing information; and carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

In this way, content information and interaction behavior information which can effectively sense real-time browsing conditions are detected in the terminal, analysis processing is carried out to obtain target browsing information, candidate recommended content which is primarily analyzed by the server is further adjusted in the terminal to obtain target recommended content for content recommendation, collaborative cooperation of the terminal and the server is achieved, content recommendation is jointly completed, accuracy of the target recommended content is high, the target recommended content can be very close to real-time real requirements of browsing objects using the terminal, and content recommendation effects are effectively improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the description of the embodiments will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 shows a schematic diagram of a system to which embodiments of the application may be applied.

Fig. 2 shows a flow chart of a content recommendation method according to an embodiment of the application.

Fig. 3 shows a flowchart of a content recommendation method according to another embodiment of the present application.

Fig. 4 shows a schematic diagram of the adjustment of content according to an embodiment of the application.

FIG. 5 illustrates a system flow diagram of a content recommendation system in a scenario.

Fig. 6 shows a frame diagram of a content recommendation system in a scenario.

Fig. 7 shows a block diagram of a content recommendation device according to another embodiment of the present application.

Fig. 8 shows a block diagram of a content recommendation device according to another embodiment of the present application.

Fig. 9 shows a block diagram of an electronic device according to an embodiment of the application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to fall within the scope of the application.

It will be appreciated that in the specific embodiment of the present application, related data such as content information and interaction information is related, and when the above embodiments of the present application are applied to specific products or technologies, user permission or consent is required to be obtained, and the collection, use and processing of related data is required to comply with related laws and regulations and standards of related countries and regions.

Fig. 1 shows a schematic diagram of a system 100 in which embodiments of the application may be applied. As shown in fig. 1, the system 100 may include a server 101 and a terminal 102.

The server 101 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, basic cloud computing services such as big data and artificial intelligence platforms, and the like.

The terminal 102 may be any device, and the terminal 102 includes, but is not limited to, a cell phone, a computer, a smart voice interaction device, a smart home appliance, a vehicle terminal, a VR/AR device, a smart watch, a computer, and the like. In some embodiments, the terminal 102 and the server 101 may be nodes in a blockchain network.

In one implementation of this example, the terminal 102 may: detecting content information corresponding to local browsing content, and detecting interaction behavior information corresponding to local content browsing interaction behavior; acquiring candidate recommended contents issued by a server 101, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request; analyzing and processing based on the content information and the interactive behavior information to obtain target browsing information; and carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

In one implementation of the present example, the server 101 may: receiving a content browsing request sent by the terminal 102; responding to the content browsing request to perform analysis processing to obtain candidate recommended content; and sending the candidate recommended content to the terminal so that the terminal can adjust the candidate recommended content according to target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation, and the target browsing information is obtained by analyzing and processing the terminal based on the interaction behavior information corresponding to the detected content browsing interaction behavior and the content information corresponding to the browsing content.

In some embodiments, the candidate recommended content and the target recommended content may be used for content pushing in the map internet of vehicles platform.

Fig. 2 schematically shows a flow chart of a content recommendation method according to an embodiment of the application. The execution subject of the content recommendation method may be any terminal, such as the terminal 102 shown in fig. 1.

As shown in fig. 2, the content recommendation method may include steps S210 to S240.

Step S210, detecting content information corresponding to local browsing content, and detecting interaction behavior information corresponding to local content browsing interaction behavior; step S220, candidate recommended contents issued by a server are obtained, wherein the candidate recommended contents are obtained by the server through analysis processing in response to a local content browsing request; step S230, analyzing and processing based on the content information and the interaction behavior information to obtain target browsing information; step S240, adjusting the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

The candidate recommended content includes candidate recommended content, the recommended content may be index content of user browsing content (such as short video or graphics context, etc.), the candidate recommended content may be information stream formed by ordering the recommended content. The candidate recommended content is determined by the server according to specific information, wherein the specific information is information (such as clicking, playing, content without clicking or playing and the like) in a specific time period in the terminal, and the candidate recommended content can be determined by the server in response to a content browsing request sent by the terminal and sent to the terminal.

The browsing contents are contents locally browsed by using a browsing object (such as a specific user) of the terminal, and the content information is information describing the browsing contents, such as classification, tags, and topics of the browsing contents. The content browsing interaction behavior is various interaction behaviors of browsing objects when browsing the content, and the interaction behavior information is information describing the content browsing interaction behavior. The terminal can timely detect the content information corresponding to the local browsing content and the interactive behavior information corresponding to the content browsing interactive behavior in real time.

The content information and the interactive behavior information can accurately and effectively reflect the real-time browsing condition in the terminal, the target browsing information is obtained through analysis and processing, and the target browsing information can accurately and effectively reflect the real-time browsing requirement in the terminal. And then, the candidate recommended content is adjusted according to the target browsing information to obtain target recommended content, and the target recommended content can be very close to the real-time real demand of the browsing object of the using terminal.

In this way, based on the steps S210 to S240, the candidate recommended content is obtained through the preliminary analysis of the server, the content information and the interaction behavior information which can effectively sense the real-time browsing condition are detected in the terminal, the analysis processing is performed to obtain the target browsing information, the candidate recommended content is further adjusted in the terminal to obtain the target recommended content for content recommendation, the collaborative coordination of the terminal and the server is realized, the content recommendation is jointly completed, the accuracy of the target recommended content is high, and the target recommended content can be very close to the real-time real requirement of the browsing object using the terminal, so that the content recommendation effect is effectively improved.

An embodiment corresponding to the specific procedure of each step performed when content recommendation is performed in the embodiment of fig. 2 is described below.

In step S210, content information corresponding to the local browsing content is detected, and interaction behavior information corresponding to the local content browsing interaction behavior is detected.

In the process that the browsing object uses the terminal to browse the content, the terminal can detect the content information of the browsed content, and the content information can comprise information such as classification, labels, themes and the like of the browsed content. The browsing object can interact with the terminal in the process of browsing the content to generate content browsing interaction behavior, and the terminal can detect the content browsing interaction behavior to obtain corresponding interaction behavior information.

The content information can reflect the real-time browsing condition in the terminal from the aspect of the content characteristics, and the interactive behavior information can reflect the real-time browsing condition in the terminal from the aspect of the behavior action characteristics, so that the real-time browsing condition in the terminal can be accurately and effectively reflected in combination.

In one embodiment, the content browsing interaction behavior includes a content exposure behavior and a browsing interaction behavior of a browsing object for browsing content; in step S210, detecting interaction behavior information corresponding to the local content browsing interaction behavior includes:

Detecting local content exposure behavior to obtain content exposure behavior information; detecting local browsing interaction behaviors to obtain browsing interaction behavior information; and taking the content exposure behavior information and the browsing interaction behavior information as interaction behavior information.

Content exposure behavior, i.e., related exposure behavior actions of the content, such as actions of exposure, scrolling, evaluation, etc., content exposure behavior information, i.e., descriptive information of related exposure behavior actions, such as exposure duration, scrolling speed, number of exposures, number of scrolling, and content evaluation (e.g., uncomfortable picture, poor quality, title party, etc.), and each action is distant from the current time interval, etc.

The browsing interaction behavior is related browsing behavior actions of content detail pages in the terminal, such as behavior actions of collection, stay, attention position, forwarding and the like, and the browsing interaction behavior information is descriptive information of related browsing behavior actions, such as whether collection, stay time, collection, attention position is an entering comment, forwarding and the like.

The content exposure behavior information and the browsing interaction behavior information are used as interaction behavior information, and the real-time browsing condition in the terminal can be accurately and effectively reflected from two angles of the content exposure behavior and the browsing interaction behavior.

In another embodiment, the content browsing interaction behavior includes only content exposure behavior of the browsing object for browsing content; in step S210, detecting interaction behavior information corresponding to the local content browsing interaction behavior includes: detecting local content exposure behavior to obtain content exposure behavior information; the content exposure behavior information is used as interaction behavior information.

In one embodiment, the content exposure behavior includes a positive feedback content exposure behavior and a negative feedback content exposure behavior; detecting local content exposure behavior to obtain content exposure behavior information, including:

detecting local positive feedback content exposure behavior to obtain positive feedback interaction information; detecting local negative feedback content exposure behavior to obtain negative feedback interaction information; and taking the positive feedback interaction information and the negative feedback interaction information as content exposure behavior characteristic information.

The positive feedback content exposure behavior is a content exposure behavior with a browsing status positive feedback effect, such as clicking, sliding browsing or playing content. The positive feedback interaction information is behavior information related to positive feedback content exposure behavior. Negative feedback content exposure behavior is content exposure behavior that does not have a browsing status positive feedback effect, such as exposure without clicking, sliding browsing, and too fast and playing content without praise, etc.

Compared with the mode that only the positive feedback interaction information is used as the content exposure behavior characteristic information in the related technology, the method can further accurately and effectively reflect the real-time browsing condition.

In one embodiment, the browsing interaction behavior includes a true browsing interaction behavior and a pseudo browsing interaction behavior; detecting local browsing interaction behavior to obtain browsing interaction behavior information comprises the following steps:

detecting a local real browsing interaction behavior to obtain real browsing interaction behavior information; detecting local pseudo-browsing interaction behavior to obtain pseudo-browsing interaction behavior information; and taking the real browsing interaction behavior information and the pseudo browsing interaction behavior information as browsing interaction behavior information.

The real browsing interaction behavior is a real browsing interaction behavior, for example, after clicking the browsing content, the behavior actions such as collection, stay, attention position, forwarding and the like are truly performed on the content detail page. The real browsing interaction behavior information is descriptive information of the real browsing interaction behavior.

Pseudo-browsing interactions are pseudo-browsing interactions, such as actions that title party content is misclicked into browsing and then quickly exits from browsing without sliding and browsing, or actions that after entering browsing, quickly interacts to the bottom of the content to view comments instead of viewing content descriptions in sequence.

The real browsing interaction behavior information and the pseudo browsing interaction behavior information are used as the browsing interaction behavior information, and compared with the mode that the real browsing interaction behavior information is used as the browsing interaction behavior information in the related technology, the real-time browsing condition can be further accurately and effectively reflected.

In step S220, candidate recommended content issued by the server is obtained, where the candidate recommended content is obtained by the server by performing analysis processing in response to a local content browsing request.

The candidate recommended content is determined by the server according to specific information, and the specific information is usually information in a specific time period in the terminal (such as information of clicking, playing and not clicking or playing content in the specific time period). The server can analyze the specific information based on the content prediction model to obtain evaluation results such as estimated click rate of the user browsing content, score the recommended content corresponding to the user browsing content according to the evaluation results and sort the recommended content according to a preset sorting strategy to obtain candidate recommended content.

The terminal can trigger a content browsing request through various behavior actions (such as searching and the like), after receiving the content browsing request, the server can analyze and process specific information corresponding to the content browsing request (such as specific information corresponding to account information, terminal information and the like in the content browsing request or specific information carried by the content browsing request and the like in the server) to obtain candidate recommended content, and send the candidate recommended content to the terminal.

In one embodiment, after the server issues the candidate recommended content to the terminal, before the terminal adjusts the candidate recommended content according to the target browsing information, the terminal may further: and recommending the recommended content in the candidate recommended content according to a preset recommendation strategy corresponding to the candidate recommended content.

After the candidate recommended content issued by the server is obtained from the terminal, the candidate recommended content may be recommended according to a predetermined recommendation policy, for example, the predetermined recommendation policy may be a predetermined ranking of recommended content included in the candidate recommended content, and the terminal may perform display of the recommended content according to the predetermined ranking to recommend the recommended content to the browsing object.

In step S230, analysis processing is performed based on the content information and the interaction information, so as to obtain target browsing information.

The terminal side can adopt a feature model based on deep learning or machine learning to extract the features of the content information and the interaction behavior information, and can adopt an analysis model (such as a deep interest model Deep Interest Network (DIN)) based on deep learning or machine learning to predict the extracted features so as to obtain target browsing information, such as browsing probability of different contents or preference degree of different contents. The target browsing information can accurately and effectively reflect the real-time browsing requirement in the terminal.

In one embodiment, step S230 of performing analysis processing based on the content information and the interaction behavior information to obtain target browsing information includes: performing feature extraction processing on the content information to obtain content features; performing feature extraction processing on the interactive behavior information to obtain interactive behavior features; and analyzing and processing based on the interactive behavior characteristics and the content characteristics to obtain the target browsing information.

The content features are feature vectors representing the content information, the interactive behavior features are feature vectors representing the interactive behavior information, and the terminal side can perform calculation processing based on the interactive behavior features and the content features, so that analysis processing is performed to obtain target browsing information.

In one embodiment, the interactive behavior information includes content exposure behavior information and browsing interactive behavior information, and the interactive behavior features include exposure behavior features and browsing behavior features; feature extraction processing is carried out on the interaction behavior information to obtain interaction behavior features, and the method comprises the following steps: performing feature extraction processing on the content exposure behavior information to obtain exposure behavior features; and carrying out feature extraction processing on the browsing interaction behavior information to obtain browsing behavior features.

In this embodiment, the interactive behavior information includes content exposure behavior information and browsing interactive behavior information. The content exposure behavior information is an information sequence in which behavior corresponding information is serially connected in time series according to the occurrence of behavior. The browsing interaction behavior information is an information sequence, and behavior action corresponding information in the information sequence is serially connected according to the occurrence time sequence of the behavior actions.

And respectively carrying out feature extraction processing on the content exposure behavior information and the browsing interaction behavior information to respectively obtain exposure behavior features and browsing behavior features, wherein compared with the mode of extracting the interaction behavior features after serializing and merging the content exposure behavior information and the browsing interaction behavior information according to the time sequence in the related technology, the method has the advantages that when the interaction behavior features including the exposure behavior features and the browsing behavior features are used for analyzing target browsing information, dominant analysis logic of the content exposure behavior information with intensive behavior can be avoided during analysis, and the analysis accuracy of the target browsing information is further improved.

In one embodiment, the analyzing and processing are performed based on the interactive behavior feature and the content feature to obtain the target browsing information, including:

selecting contents to be recommended from the candidate recommended contents, and extracting characteristics of the contents to be recommended corresponding to the contents to be recommended; fusing the interactive behavior features and the content features to obtain fused features; performing mutual fusion coding processing on the interactive behavior characteristics, the content characteristics to be recommended and the fusion characteristics to obtain characteristics to be identified; and carrying out prediction processing based on the features to be identified to obtain target browsing information.

The content to be recommended is content in the candidate recommended content, in one example, the content to be recommended is all content in the candidate recommended content, in one example, the content to be recommended is content in the candidate recommended content which is not browsed by the browsed object.

After obtaining the content information to be recommended (such as classification, label and theme information of the content to be recommended) to be recommended, performing feature extraction processing to obtain corresponding content features to be recommended, wherein the content features to be recommended are feature vectors representing the content information to be recommended.

And the interactive behavior features and the content features can be fused by adding or multiplying and other fusion modes to obtain fusion features. The interactive behavior feature, the content feature to be recommended and the fusion feature can be processed by means of mutual fusion coding through self-attention arithmetic processing or convolution and the like to obtain the feature to be identified, for example, the interactive behavior feature, the content feature to be recommended and the fusion feature can be input into a cyclic neural network (Recurrent Neural Network, RNN), a Long Short-Term Memory (LSTM), a gate-control cyclic unit (GRU) or a attention model (Transform) and the like to be processed by mutual fusion coding to obtain the feature to be identified.

The features of the content to be recommended and the information reflecting the real-time browsing condition are fused in the features to be identified, and the features to be identified are predicted by adopting an analysis model based on deep learning or machine learning, so that accurate target browsing information can be obtained.

In one embodiment, performing mutual fusion encoding processing on the interactive behavior feature, the content feature to be recommended and the fusion feature to obtain the feature to be identified, including: and taking the content features to be recommended as query features, the interaction behavior features as key features and the fusion features as value features to perform self-attention operation processing to obtain the features to be identified.

In this embodiment, the content feature to be recommended is used as a query feature (query), the interaction behavior feature is used as a key feature (key), and the fusion feature is used as a value feature (value), and the query feature (query), the key feature (key), and the value feature (value) are subjected to self-attention operation (self-attention), so that the feature to be identified which effectively characterizes long-term semantics can be obtained, for example, the query feature (query), the key feature (key), and the value feature (value) can be subjected to self-attention operation (self-attention) in an attention model (Transform), so that the feature to be identified can be obtained. The target browsing information is predicted based on the features to be identified, so that the accuracy of the target browsing information can be improved compared with other modes, and the content recommendation effect is further improved effectively as a whole.

In step S240, the candidate recommended content is adjusted according to the target browsing information, so as to obtain a target recommended content, which is used for content recommendation.

The target browsing information can accurately and effectively reflect the real-time browsing requirement in the terminal, and the target browsing information can be, for example, browsing probabilities of different contents or preference degrees of different contents. And according to the target browsing information, the candidate recommended content decided by the server can be adjusted, for example, the sequence of the recommended content in the candidate recommended content is adjusted, so that the target recommended content is obtained.

The target recommended content obtained through adjustment processing can effectively match with real-time browsing requirements in the terminal, and the target recommended content is used for effectively improving content recommendation effects when content recommendation is carried out.

In one embodiment, the adjusting the candidate recommended content according to the target browsing information to obtain the target recommended content includes:

determining a dynamic recommendation strategy of recommended content in the candidate recommended content according to the target browsing information; and adjusting the recommended content in the candidate recommended content according to the dynamic recommended strategy to obtain the target recommended content.

The target browsing information, such as browsing probability of different contents or preference degree of different contents, can determine a new ranking corresponding to the recommended contents in the candidate recommended contents or whether to reject some dynamic recommendation strategies such as the recommended contents from the candidate recommended contents according to the target browsing information such as browsing probability of non-contents or preference degree of different contents.

And then, according to the dynamic recommendation strategy, the candidate recommended content is subjected to adjustment processing, and the recommended content in the candidate recommended content can be subjected to adjustment such as reordering or removing some recommended content, so as to obtain target recommended content.

In one embodiment, the adjusting the candidate recommended content according to the dynamic recommendation policy to obtain the target recommended content includes: determining recommended contents to be browsed from recommended contents included in the candidate recommended contents; and adjusting the recommended content to be browsed according to the dynamic recommendation strategy to obtain the target recommended content.

The terminal can display the candidate recommended content issued by the server locally after acquiring the candidate recommended content, the browsing object can browse the recommended content in the candidate recommended content, and the user can browse the content deeply by accessing the content detail page through the recommended content. The recommended content to be browsed, namely the recommended content to be browsed of the browsing object in the candidate recommended content, can be unexposed recommended content in the candidate recommended content. In one example, referring to fig. 4, the candidate recommended contents include recommended contents a, b, c, d and e arranged in sequence, and if a and b of the candidate recommended contents are exposed, the unexposed recommended contents c, d and e are recommended contents to be browsed.

And adjusting the recommended content to be browsed in the candidate recommended content according to the dynamic recommended strategy, and re-sequencing the recommended content to be browsed, removing certain recommended content or inserting the recommended content to be browsed, and the like to obtain target recommended content. In one example, referring to fig. 4, the recommended content (c, d, and e) to be browsed is adjusted to obtain the target recommended content (d, e, and c).

Further, after step S240, the terminal may recommend the target recommended content, and output and display the recommended content in the target recommended content according to the order when the target recommended content is recommended.

Fig. 3 schematically shows a flow chart of a content recommendation method according to another embodiment of the application. The execution subject of the content recommendation method may be any server, such as the server 101 shown in fig. 1.

As shown in fig. 3, the content recommendation method may include steps S310 to S330:

step S310, receiving a content browsing request sent by a terminal; step S320, analyzing and processing are carried out in response to the content browsing request, and candidate recommended content is obtained; step S330, the candidate recommended content is issued to the terminal, so that the terminal adjusts the candidate recommended content according to target browsing information to obtain target recommended content, the target recommended content is used for content recommendation, and the target browsing information is obtained by analyzing and processing the terminal based on the detected interactive behavior information corresponding to the content browsing interactive behavior and the content information corresponding to the browsing content.

The candidate recommended content is the candidate recommended content, the candidate recommended content is determined by a server according to specific information, the specific information is usually information (such as clicked, played and content without clicked or played) in a specific time period in the terminal, and the server can respond to a content browsing request sent by the terminal to determine the candidate recommended content and send the candidate recommended content to the terminal.

The content information and the interactive behavior information can accurately and effectively reflect the real-time browsing condition in the terminal, the target browsing information is obtained in the terminal through analysis and processing, and the target browsing information can accurately and effectively reflect the real-time browsing requirement in the terminal. And furthermore, the terminal adjusts the candidate recommended content according to the target browsing information to obtain the target recommended content, and the target recommended content can be very close to the real-time real demand of the browsing object using the terminal.

In this way, based on the steps S310 to S330, the candidate recommended content is obtained through the preliminary analysis of the server, the content information and the interaction behavior information which can effectively sense the real-time browsing condition are detected in the terminal, the analysis processing is performed to obtain the target browsing information, the candidate recommended content is further adjusted in the terminal to obtain the target recommended content for content recommendation, the collaborative coordination of the terminal and the server is realized, the content recommendation is jointly completed, the accuracy of the target recommended content is high, the target recommended content can be very close to the real-time real requirement of the browsing object using the terminal, and the content recommendation effect is further effectively improved.

The foregoing embodiments are further described below in connection with a process of content recommendation in an application scenario. Referring to fig. 5 and 6, fig. 5 shows a system flowchart in the content recommendation system in this scenario. Fig. 6 shows a system architecture diagram for content recommendation in this scenario.

Referring to fig. 5, the content recommendation system may include a content production end, a content consumption end, and a cloud end (i.e., a cloud server). The cloud may include: the uplink and downlink content interface service, the content outlet service, the content database, the dispatch center service, the manual auditing system, the duplication eliminating service, the statistical reporting interface service, the statistical database, the recommended cloud recall system, the recommended cloud sorting service and the content consumption end can comprise: terminal behavior collection and feature extraction system, terminal remixing and arranging system and terminal caching system. In this scenario, steps 5.1.1 to 5.1.12 and steps 5.2.1 to 5.2.5 may be implemented in the content recommendation system, and it may be understood that the order of these steps may be adjusted according to actual requirements.

Step 5.1.1, uploading the released content, where the content producer may release the content to the uplink and downlink content interface service, where the content producer may be a PGC (Professional Generated Content, professional production content), UGC (User Generated Content ), MCN (Multi-Channel Network), or PUGC (Professional Generated Content + User Generated Content, i.e. "professional user produced content" or "expert produced content"), and the content producer may release, through a mobile terminal or a back-end interface system, user browsing content provided by a local or Network release system, including graphics context, short video, small video, etc., where the released user browsing content may be used as a message source (Feeds) in a recommendation system. The content production end can firstly acquire the interface address of the uploading server through communication with the uplink and downlink content interface service, and then browse the content at the publishing user.

And 5.1.2, storing the source file, and storing the source file of the content browsed by the user to a content storage service by an uplink and downlink content interface service. And 5.1.3, writing meta information and content, wherein the uplink and downlink content interface service writes the meta information of the content browsed by the user and the content browsed by the user into a content database, wherein the meta information comprises classification (for example, one, two and three categories of classification and label information, such as an article, one category is science and technology, two categories of classification are smart phones, three categories of classification are domestic phones, and label information is manufacturer) and the like of the content browsed by the user in the processes of size, cover map link, title, release time, account author, source channel, warehouse-in time and manual auditing. And 5.1.4, the content enters the dispatching center, and the uplink and downlink content interface service transmits the browsed content of the user to the dispatching center service. And 5.1.5, the weight removing result is obtained, and the dispatching center service dispatches the weight removing service to mark and filter the repeatedly-warehoused user browsing content. And 5.1.6, manually checking, wherein the dispatching center service dispatches the manual checking system to manually check the browsed content of the user which cannot be processed by the machine, the information in the content database can be read in the manual checking process, and meanwhile, the result and the state of the manual checking can be returned to the content database for storage. And 5.1.7, reading the meta information, and recommending the cloud recall system to acquire the meta information from the content database. And 5.1.8, updating the meta-information, and updating the meta-information in the content database by the service of the dispatching center according to the duplication eliminating result and the like.

Further, referring to fig. 6, in this scenario, the terminal may include a detection module, an analysis module, a request decision module, an adjustment module, and a display module, and the cloud may include a receiving module, a decision module, and a sending module. The terminal may be the content consumer in fig. 5.

In the cloud, the receiving module can receive a content browsing request sent by a terminal (namely a content consumption end); the decision module can respond to the content browsing request to carry out analysis processing so as to obtain candidate recommended content; the sending module can send the candidate recommended content to the terminal so that the terminal can adjust the candidate recommended content according to target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation, and the target browsing information is obtained by analyzing and processing the terminal based on the interaction behavior information corresponding to the detected content browsing interaction behavior and the content information corresponding to the browsing content.

The decision module may include a recommended cloud recall system and a recommended cloud sequencing service shown in fig. 6, and the decision module may specifically include steps 5.1.9 to 5.1.10, where the decision module performs analysis processing in response to a content browsing request.

And 5.1.9, recalling by a recommendation algorithm, and analyzing specific information based on a content prediction model in a recommendation cloud recall system to obtain scores such as estimated click rate of content browsed by a user. In the recommended cloud recall system, a recommendation algorithm can be adopted to recall user browsing content corresponding to meta information in response to a content browsing request of a terminal (namely a content consumption end). In step 5.1.10, recall results are aggregated, and the recommended cloud ranking service may score recommended content corresponding to the browsed content of the user according to the evaluation result and rank the recommended content according to a predetermined ranking policy to obtain candidate recommended content, where the predetermined ranking policy may be a service rule policy (such as a duty ratio of the image-text content and the video content or an operation top-set and insert content).

The sending module may include a content distribution outlet service that issues candidate recommended content to a terminal (i.e., content consumer) based on the diversity order distribution, step 5.1.11.

In the terminal, the detection module can detect the content information corresponding to the local browsing content and detect the interactive behavior information corresponding to the local content browsing interactive behavior; the acquisition module can acquire candidate recommended contents issued by the server, wherein the candidate recommended contents are obtained by the server by responding to a local content browsing request and analyzing and processing; the analysis module can analyze and process based on the content information and the interaction behavior information to obtain target browsing information; the adjustment module can perform adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation. The request decision module may be used to trigger a content browsing request. The presentation module may present the target recommended content for recommendation to the user.

The acquisition module acquires candidate recommended content issued by the server, namely, step 5.2.1, acquires a cloud content index and content, and the content consumption end acquires the candidate recommended content issued by the cloud, and based on the user browsing content (such as user browsing content in a content database) in the candidate recommended content consumption end, the user browsing content in the browsing cloud can be accessed.

The candidate recommended content is obtained by analyzing and processing a server (in this scenario, a cloud) in response to a local content browsing request, the candidate recommended content is specifically determined according to specific information, and the process of the cloud obtaining the specific information to make a decision may include steps 5.2.2 to 5.2.5.

And 5.2.2, reporting the message information, wherein the uplink and downlink content interface service reports the message information of each account to the statistical interface server, and the message information can comprise message time, content type and the like. And 5.2.3, reporting the statistical information, and reporting the content information in the terminal to a statistical reporting interface service by the content consumption terminal. And 5.2.4, writing statistical information, and writing specific information comprising the text information and the content information into a statistical database by the statistical reporting interface service. And 5.2.5, reading the statistical information to recommend contents, and reading the specific information from the statistical database by the recommendation cloud recall system to analyze and process to obtain candidate recommended contents.

The terminal behavior collection and feature extraction system in the terminal can comprise a detection module and an analysis module, and the terminal remixing and arranging system can comprise an adjustment module.

Further, the content consumption end can always recommend the recommended content in the candidate recommended content according to a preset recommendation strategy corresponding to the candidate recommended content.

The content browsing interaction behavior comprises a content exposure behavior and a browsing interaction behavior of a browsing object aiming at browsing content; detecting interaction behavior information corresponding to local content browsing interaction behavior, including: detecting local content exposure behavior to obtain content exposure behavior information; detecting local browsing interaction behaviors to obtain browsing interaction behavior information; and taking the content exposure behavior information and the browsing interaction behavior information as interaction behavior information.

The content exposure behavior comprises positive feedback content exposure behavior and negative feedback content exposure behavior; detecting local content exposure behavior to obtain content exposure behavior information, including: detecting local positive feedback content exposure behavior to obtain positive feedback interaction information; detecting local negative feedback content exposure behavior to obtain negative feedback interaction information; and taking the positive feedback interaction information and the negative feedback interaction information as content exposure behavior characteristic information. The browsing interaction behavior comprises a true browsing interaction behavior and a false browsing interaction behavior; detecting local browsing interaction behavior to obtain browsing interaction behavior information comprises the following steps: detecting a local real browsing interaction behavior to obtain real browsing interaction behavior information; detecting local pseudo-browsing interaction behavior to obtain pseudo-browsing interaction behavior information; and taking the real browsing interaction behavior information and the pseudo browsing interaction behavior information as browsing interaction behavior information.

Analyzing and processing based on the content information and the interaction behavior information to obtain target browsing information, wherein the method comprises the following steps: performing feature extraction processing on the content information to obtain content features; performing feature extraction processing on the interactive behavior information to obtain interactive behavior features; and analyzing and processing based on the interactive behavior characteristics and the content characteristics to obtain the target browsing information.

The interactive behavior features comprise exposure behavior features and browsing behavior features; feature extraction processing is carried out on the interaction behavior information to obtain interaction behavior features, and the method comprises the following steps: performing feature extraction processing on the content exposure behavior information to obtain exposure behavior features; and carrying out feature extraction processing on the browsing interaction behavior information to obtain browsing behavior features. The terminal cache system in the content consumption end can cache the content features and the interaction behavior features extracted from the analysis module.

Analyzing and processing based on the interactive behavior characteristics and the content characteristics to obtain target browsing information, wherein the method comprises the following steps: selecting contents to be recommended from the candidate recommended contents, and extracting characteristics of the contents to be recommended corresponding to the contents to be recommended; fusing the interactive behavior features and the content features to obtain fused features; performing mutual fusion coding processing on the interactive behavior characteristics, the content characteristics to be recommended and the fusion characteristics to obtain characteristics to be identified; and carrying out prediction processing based on the features to be identified to obtain target browsing information.

Performing mutual fusion coding processing on the interactive behavior characteristic, the content characteristic to be recommended and the fusion characteristic to obtain a characteristic to be identified, wherein the method comprises the following steps: and taking the content features to be recommended as query features, the interaction behavior features as key features and the fusion features as value features to perform self-attention operation processing to obtain the features to be identified.

Adjusting the candidate recommended content according to the target browsing information to obtain the target recommended content, including: determining a dynamic recommendation strategy of recommended content in the candidate recommended content according to the target browsing information; and adjusting the recommended content in the candidate recommended content according to the dynamic recommended strategy to obtain the target recommended content. The candidate recommended content is adjusted according to the dynamic recommended strategy to obtain target recommended content, which comprises the following steps: determining recommended contents to be browsed from recommended contents included in the candidate recommended contents; and adjusting the recommended content to be browsed according to the dynamic recommendation strategy to obtain the target recommended content.

In this way, by applying the foregoing embodiment of the present application in this scenario, at least the following beneficial effects are provided, in which the content consumption end detects the content information and the interaction behavior information that can effectively sense the real-time browsing condition, and performs analysis processing to obtain the target browsing information, in which the content consumption end further adjusts the candidate recommended content primarily analyzed by the cloud to obtain the target recommended content for performing content recommendation, so as to implement collaborative cooperation between the content consumption end and the cloud, complete content recommendation together, and the target recommended content has high accuracy, and can be very close to the real-time real demand of the browsing object using the content consumption end, and effectively improve the content recommendation effect.

In order to facilitate better implementation of the content recommendation method provided by the embodiment of the application, the embodiment of the application also provides a content recommendation device based on the content recommendation method. Where the meaning of the terms is the same as in the content recommendation method described above, specific implementation details may be referred to in the description of the method embodiments. Fig. 7 shows a block diagram of a content recommendation device according to an embodiment of the present application. Fig. 8 shows a block diagram of a content recommendation device according to another embodiment of the present application.

As shown in fig. 7, the content recommendation device 600 is applied to a terminal, and the content recommendation device 600 may include a detection module 610, an acquisition module 620, an analysis module 630, and an adjustment module 640.

The detection module 610 may be configured to detect content information corresponding to local browsing content, and detect interaction behavior information corresponding to local content browsing interaction behavior; the obtaining module 620 may be configured to obtain candidate recommended content issued by a server, where the candidate recommended content is obtained by analyzing and processing the server in response to a local content browsing request; the analysis module 630 may be configured to perform analysis processing based on the content information and the interaction behavior information to obtain target browsing information; the adjustment module 640 may be configured to perform adjustment processing on the candidate recommended content according to the target browsing information, so as to obtain a target recommended content, where the target recommended content is used for content recommendation.

As shown in fig. 8, the content recommendation device 700 is applied to a server, and the content recommendation device 700 may include a receiving module 710, a decision module 720, and a transmitting module 730.

The receiving module 710 may be configured to receive a content browsing request sent by a terminal; the decision module 720 may be configured to perform analysis processing in response to the content browsing request, so as to obtain candidate recommended content; the sending module 730 may be configured to send the candidate recommended content to the terminal, so that the terminal adjusts the candidate recommended content according to target browsing information to obtain target recommended content, where the target recommended content is used for content recommendation, and the target browsing information is obtained by analyzing and processing the terminal based on interaction behavior information corresponding to the detected content browsing interaction behavior and content information corresponding to the browsed content.

It should be noted that although in the above detailed description several modules or units of a device for action execution are mentioned, such a division is not mandatory. Indeed, the features and functions of two or more modules or units described above may be embodied in one module or unit in accordance with embodiments of the application. Conversely, the features and functions of one module or unit described above may be further divided into a plurality of modules or units to be embodied.

In addition, the embodiment of the present application further provides an electronic device, which may be a terminal or a server, as shown in fig. 9, which shows a schematic structural diagram of the electronic device according to the embodiment of the present application, specifically:

the electronic device may include one or more processing cores 'processors 801, one or more computer-readable storage media's memory 802, power supply 803, and input unit 804, among other components. It will be appreciated by those skilled in the art that the electronic device structure shown in fig. 9 is not limiting of the electronic device and may include more or fewer components than shown, or may combine certain components, or a different arrangement of components. Wherein:

the processor 801 is a control center of the electronic device, connects various parts of the entire computer device using various interfaces and lines, and performs various functions of the computer device and processes data by running or executing software programs and/or modules stored in the memory 802, and calling data stored in the memory 802, thereby detecting the electronic device. Optionally, the processor 801 may include one or more processing cores; preferably, the processor 801 may integrate an application processor that primarily handles operating systems, user pages, applications, etc., with a modem processor that primarily handles wireless communications. It will be appreciated that the modem processor described above may not be integrated into the processor 801.

The memory 802 may be used to store software programs and modules, and the processor 801 executes various functional applications and data processing by executing the software programs and modules stored in the memory 802. The memory 802 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program (such as a sound playing function, an image playing function, etc.) required for at least one function, and the like; the storage data area may store data created according to the use of the computer device, etc. In addition, memory 802 may include high-speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid-state storage device. Accordingly, the memory 802 may also include a memory controller to provide the processor 801 with access to the memory 802.

The electronic device further comprises a power supply 803 for powering the various components, preferably the power supply 803 can be logically coupled to the processor 801 via a power management system such that functions such as managing charging, discharging, and power consumption are performed by the power management system. The power supply 803 may also include one or more of any components, such as a direct current or alternating current power supply, a recharging system, a power failure detection circuit, a power converter or inverter, a power status indicator, and the like.

The electronic device may further comprise an input unit 804, which input unit 804 may be used for receiving input digital or character information and for generating keyboard, mouse, joystick, optical or trackball signal inputs in connection with user settings and function control.

Although not shown, the electronic device may further include a display unit or the like, which is not described herein. In particular, in this embodiment, the processor 801 in the electronic device loads executable files corresponding to the processes of one or more computer programs into the memory 802 according to the following instructions, and the processor 801 executes the computer programs stored in the memory 802, so as to implement the functions of the foregoing embodiments of the present application.

In one embodiment, the processor 801 may perform: detecting content information corresponding to local browsing content, and detecting interaction behavior information corresponding to local content browsing interaction behavior; acquiring candidate recommended contents issued by a server, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request; analyzing and processing based on the content information and the interactive behavior information to obtain target browsing information; and carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

In another embodiment, for example, the processor 801 may perform: receiving a content browsing request sent by a terminal; responding to the content browsing request to perform analysis processing to obtain candidate recommended content; and sending the candidate recommended content to the terminal so that the terminal can adjust the candidate recommended content according to target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation, and the target browsing information is obtained by analyzing and processing the terminal based on the interaction behavior information corresponding to the detected content browsing interaction behavior and the content information corresponding to the browsing content.

It will be appreciated by those of ordinary skill in the art that all or part of the steps of the various methods of the above embodiments may be performed by a computer program, or by computer program control related hardware, which may be stored in a computer readable storage medium and loaded and executed by a processor.

To this end, embodiments of the present application also provide a computer readable storage medium having stored therein a computer program that can be loaded by a processor to perform the steps of any of the methods provided by the embodiments of the present application.

Wherein the computer-readable storage medium may comprise: read Only Memory (ROM), random access Memory (RAM, random Access Memory), magnetic or optical disk, and the like.

Since the computer program stored in the computer readable storage medium may execute the steps of any one of the methods provided in the embodiments of the present application, the beneficial effects that can be achieved by the methods provided in the embodiments of the present application may be achieved, which are detailed in the previous embodiments and are not described herein.

According to one aspect of the present application, there is provided a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions to cause the computer device to perform the methods provided in the various alternative implementations of the application described above.

Other embodiments of the application will be apparent to those skilled in the art from consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of the application following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the application pertains.

It will be understood that the application is not limited to the embodiments which have been described above and shown in the drawings, but that various modifications and changes can be made without departing from the scope thereof.

Claims

1. A content recommendation method, applied to a terminal, comprising:

detecting content information corresponding to local browsing content, and detecting interaction behavior information corresponding to local content browsing interaction behavior;

acquiring candidate recommended contents issued by a server, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request;

analyzing and processing based on the content information and the interactive behavior information to obtain target browsing information;

and carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

2. The content recommendation method according to claim 1, wherein the content browsing interaction behavior includes a content exposure behavior and a browsing interaction behavior of a browsing object for browsing content; the detecting the interaction behavior information corresponding to the local content browsing interaction behavior comprises the following steps:

Detecting the local content exposure behavior to obtain content exposure behavior information;

detecting the local browsing interaction behavior to obtain browsing interaction behavior information;

and taking the content exposure behavior information and the browsing interaction behavior information as the interaction behavior information.

3. The content recommendation method according to claim 2, wherein the content exposure behavior includes a positive feedback content exposure behavior and a negative feedback content exposure behavior; the detecting the local content exposure behavior to obtain content exposure behavior information comprises the following steps:

detecting the local positive feedback content exposure behavior to obtain positive feedback interaction information;

detecting the local exposure behavior of the negative feedback content to obtain negative feedback interaction information;

and taking the positive feedback interaction information and the negative feedback interaction information as the content exposure behavior characteristic information.

4. The content recommendation method according to claim 2, wherein the browsing interaction behavior includes a true browsing interaction behavior and a pseudo browsing interaction behavior; the detecting the local browsing interaction behavior to obtain browsing interaction behavior information includes:

detecting the local real browsing interaction behavior to obtain real browsing interaction behavior information;

Detecting the local pseudo-browsing interaction behavior to obtain pseudo-browsing interaction behavior information;

and taking the real browsing interaction behavior information and the pseudo browsing interaction behavior information as the browsing interaction behavior information.

5. The content recommendation method according to claim 1, wherein the analyzing based on the content information and the interaction information to obtain target browsing information includes:

performing feature extraction processing on the content information to obtain content features;

performing feature extraction processing on the interactive behavior information to obtain interactive behavior features;

and analyzing and processing based on the interactive behavior characteristics and the content characteristics to obtain the target browsing information.

6. The content recommendation method according to claim 5, wherein the interactive behavior information includes content exposure behavior information and browsing interactive behavior information, and the interactive behavior features include exposure behavior features and browsing behavior features; the feature extraction processing is performed on the interactive behavior information to obtain interactive behavior features, including:

performing feature extraction processing on the content exposure behavior information to obtain the exposure behavior feature;

And carrying out feature extraction processing on the browsing interaction behavior information to obtain the browsing behavior features.

7. The content recommendation method according to claim 5, wherein the analyzing based on the interactive behavior feature and the content feature to obtain the target browsing information includes:

selecting content to be recommended from the candidate recommended content, and extracting content characteristics to be recommended corresponding to the content to be recommended;

carrying out fusion processing on the interactive behavior characteristics and the content characteristics to obtain fusion characteristics;

performing mutual fusion coding processing on the interactive behavior characteristics, the content characteristics to be recommended and the fusion characteristics to obtain characteristics to be identified;

and carrying out prediction processing based on the features to be identified to obtain the target browsing information.

8. The content recommendation method according to claim 7, wherein the performing the inter-fusion encoding processing on the interactive behavior feature, the content feature to be recommended, and the fusion feature to obtain the feature to be identified includes:

and performing self-attention operation processing by taking the content features to be recommended as query features, the interaction behavior features as key features and the fusion features as value features to obtain the features to be identified.

9. The content recommendation method according to claim 1, wherein the adjusting the candidate recommended content according to the target browsing information to obtain the target recommended content includes:

determining a dynamic recommendation strategy of recommended content in the candidate recommended content according to the target browsing information;

and adjusting the recommended content in the candidate recommended content according to the dynamic recommended strategy to obtain the target recommended content.

10. The content recommendation method according to claim 9, wherein the adjusting the candidate recommended content according to the dynamic recommendation policy to obtain the target recommended content includes:

determining recommended contents to be browsed from recommended contents included in the candidate recommended contents;

and adjusting the recommended content to be browsed according to the dynamic recommendation strategy to obtain the target recommended content.

11. The content recommendation method according to claim 1, wherein before the candidate recommended content is subjected to adjustment processing according to the target browsing information, the method further comprises:

and recommending the recommended content in the candidate recommended content according to a preset recommendation strategy corresponding to the candidate recommended content.

12. A content recommendation method, applied to a server, comprising:

receiving a content browsing request sent by a terminal;

responding to the content browsing request to perform analysis processing to obtain candidate recommended content;

and sending the candidate recommended content to the terminal so that the terminal can adjust the candidate recommended content according to target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation, and the target browsing information is obtained by analyzing and processing the terminal based on the interaction behavior information corresponding to the detected content browsing interaction behavior and the content information corresponding to the browsing content.

13. A content recommendation device, characterized by being applied to a terminal, comprising:

the detection module is used for detecting content information corresponding to local browsing content and detecting interaction behavior information corresponding to local content browsing interaction behavior;

the acquisition module is used for acquiring candidate recommended contents issued by the server, wherein the candidate recommended contents are obtained by analyzing and processing the server in response to a local content browsing request;

the analysis module is used for carrying out analysis processing based on the content information and the interaction behavior information to obtain target browsing information;

And the adjustment module is used for carrying out adjustment processing on the candidate recommended content according to the target browsing information to obtain target recommended content, wherein the target recommended content is used for content recommendation.

14. A computer-readable storage medium, having stored thereon a computer program which, when executed by a processor of a computer, causes the computer to perform the method of any of claims 1 to 12.

15. An electronic device, comprising: a memory storing a computer program; a processor reading the computer program stored in the memory to perform the method of any one of claims 1 to 12.

16. A computer program product, characterized in that the computer program product comprises a computer program which, when executed by a processor, implements the method of any one of claims 1 to 12.