CN110162691B

CN110162691B - Topic recommendation, operation method, device and machine equipment in online content service

Info

Publication number: CN110162691B
Application number: CN201811362047.7A
Authority: CN
Inventors: 俄万有
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2018-11-15
Filing date: 2018-11-15
Publication date: 2023-05-23
Anticipated expiration: 2038-11-15
Also published as: CN110162691A

Abstract

The invention discloses a topic recommendation method, an operation method, a video topic recommendation device and machine equipment in online video content service. The method comprises the following steps: acquiring basic information of given content and topic information corresponding to the given content; performing correlation calculation of topic information and basic information according to the basic information and the topic information, and obtaining candidate topic information from topic information corresponding to given content according to the correlation; according to group characteristic information related to the user and candidate topic information corresponding to given content, acquiring a recommended topic facing the group of the user, and distributing the recommended topic facing the group of the user. Therefore, the method is not dependent on manpower any more, topics related to the content, namely recommended topics of the given content, are automatically obtained in a personalized way, topic operation efficiency is improved, labor cost is greatly reduced, and accurate recommendation of different topics for the same content can be realized for different users.

Description

Topic recommendation, operation method, device and machine equipment in online content service

Technical Field

The invention relates to the technical field of internet application, in particular to a topic recommendation method, an operation method, a video topic recommendation device and machine equipment in online video content service.

Background

With the vigorous development of the mobile internet, the sharing and spreading of topics has become a network behavior habit of internet users. For many contents distributed in the internet, for example, a movie video content, users can continuously create and share topics related to the movie video content while watching the movie video content.

For the watched video content, users often release information related to the watched video content through various ways in the internet to participate in the discussion of the watched video content as topics, so that topic creation and sharing propagation related to the video content are realized. Users themselves and other users can participate in the discussion of the topic, so that more topics are induced, and a virtuous circle of the distributed content of the internet on the topics is formed.

That is, the content distribution by the user is oriented, and the presentation of the topics related to the content is essentially performed, so as to promote the creation and sharing propagation of the topics related to the content.

However, the topics related to the content are obtained by manually editing and editing the content, and are displayed on a topic operation page, so that the limitation of low operation efficiency and very high labor cost exists, and inaccuracy of topic operation is necessarily caused. It is difficult to achieve accurate pushing of different topics for the same content for different users, so accurate recommendation of topics related to the content for online content services is needed.

Disclosure of Invention

The invention provides a topic recommendation method, an operation method, a video topic recommendation device and machine equipment in online video content service, which can accurately recommend topics.

A method of topic recommendation in an online content service, the method comprising:

for given content of an online content service, acquiring basic information of the given content and topic information corresponding to the given content;

performing correlation calculation of the topic information and the basic information for the given content according to the basic information and the topic information, and obtaining candidate topic information from topic information corresponding to the given content according to the calculated correlation;

according to group characteristic information related to the user and candidate topic information corresponding to the given content, acquiring a recommended topic facing the group where the user is located, and distributing the recommended topic facing the group where the user is located.

A topic operation method of an online content service, the method comprising:

acquiring a content request sent by an online content service, wherein the content request carries user account information;

returning basic information and topic information of given content to the online content service according to the content request, performing correlation calculation on the basic information and topic information of the given content, and obtaining candidate topic information from the topic information according to the calculated correlation;

acquiring recommended topics which mark the group where the user is located for the user account information from the candidate topic information according to group characteristic information corresponding to the user account information carried by the content request;

and fusing the recommended topics and the manual operation topics of the given content to obtain topic data, and returning the topic data to the online content service according to the user account information.

A video topic recommendation method, the method comprising:

for given video content of an online video content service, acquiring video base information of the given video content and topic information corresponding to the given video content;

according to the video basic information and topic information, performing correlation calculation of the topic information and the basic information for the given video content, and obtaining candidate topic information from topic information corresponding to the given video content according to the calculated correlation;

Acquiring recommended topics oriented to a group where a video user is located according to group characteristic information related to the video user and candidate topic information corresponding to the given video content, and distributing the recommended topics oriented to the group where the video user is located.

A topic recommendation device in an online content service, the device comprising:

the information acquisition module is used for acquiring basic information of given content and topic information corresponding to the given content for the given content of the online content service;

the candidate acquisition module is used for executing correlation calculation of the topic information and the basic information for the given content according to the basic information and the topic information, and acquiring candidate topic information from the topic information corresponding to the given content according to the calculated correlation;

the recommendation generation module is used for acquiring recommendation topics facing the group where the user is located according to group characteristic information related to the user and candidate topic information corresponding to the given content, and distributing the recommendation topics facing the group where the user is located.

A topic operation apparatus of an online content service, the apparatus comprising:

The request acquisition module is used for acquiring a content request sent by the online content service, wherein the content request carries user account information;

a candidate calculating module, configured to return basic information and topic information of a given content to the online content service according to the content request, perform correlation calculation on the basic information and topic information of the given content, and obtain candidate topic information from the topic information according to the calculated correlation;

the recommendation acquisition execution module is used for acquiring recommendation topics which are marked by the group where the user is located and are oriented to the user account information from the candidate topic information of the given content according to group characteristic information corresponding to the user account information carried by the content request;

the fusion module is used for fusing the recommended topics and the manual operation topics of the given content to obtain topic data, and returning the topic data to the online content service according to the user account information.

A video topic recommendation device, the device comprising:

the video related information acquisition module is used for acquiring video basic information of given video content and topic information corresponding to the given video content for the given video content of the online video content service;

The candidate topic acquisition module is used for executing correlation calculation of the topic information and the basic information for the given video content according to the video basic information and the topic information, and acquiring candidate topic information from topic information corresponding to the given video content according to the calculated correlation;

the recommendation topic obtaining module is used for obtaining recommendation topics oriented to the group where the video user is located according to group characteristic information related to the video user and a plurality of candidate topic information corresponding to the given video content, and distributing the recommendation topics oriented to the group where the video user is located.

A machine apparatus, comprising:

a processor; and

a memory having stored thereon computer readable instructions which when executed by the processor implement a method as described above.

The technical scheme provided by the embodiment of the invention can comprise the following beneficial effects:

for a given content of an online content service, recommendation of related topics is realized for a user, basic information of the given content and topic information corresponding to the given content are firstly obtained, then correlation calculation of the topic information and the basic information is carried out for the given content according to the basic information and the topic information, candidate topic information is obtained from the topic information corresponding to the given content according to the calculated correlation, finally recommended topics of a group of users are obtained according to group characteristic information related to the user and the candidate topic information corresponding to the given content, and the recommended topics are distributed for the group of the users, so that the manual work is not relied, the related topics of the content, namely the recommended topics of the given content, topic operation efficiency is improved, labor cost is greatly reduced, and accurate recommendation of different topics for the same content can be realized for different users.

It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention as claimed.

Drawings

The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the invention and together with the description, serve to explain the principles of the invention.

FIG. 1 is a schematic illustration of an implementation environment in accordance with the present invention;

FIG. 2 is a block diagram of an apparatus according to an example embodiment

FIG. 3 is a flowchart illustrating a topic recommendation method in an online content service in accordance with an exemplary embodiment;

FIG. 4 is a flow chart depicting step 310, shown in accordance with the corresponding embodiment of FIG. 3;

FIG. 5 is a flow chart depicting step 330, shown in accordance with the corresponding embodiment of FIG. 3;

FIG. 6 is a flow chart depicting step 350, in accordance with the corresponding embodiment of FIG. 3;

FIG. 7 is a flow chart depicting step 351, shown in accordance with the corresponding embodiment of FIG. 6;

FIG. 8 is a flowchart illustrating a topic operation method for an online content service in accordance with an exemplary embodiment;

FIG. 9 is a flowchart illustrating a video topic recommendation method in accordance with an exemplary embodiment;

Fig. 10 is a system architecture diagram of a video topic operation system shown in accordance with an exemplary embodiment;

FIG. 11 is a system architecture diagram of the topic recommendation system shown in accordance with the corresponding embodiment of FIG. 10;

FIG. 12 is a timing diagram for video topic generation shown in accordance with the corresponding embodiment of FIG. 11;

FIG. 13 is a block diagram illustrating a topic recommendation device in an online content service in accordance with an exemplary embodiment;

FIG. 14 is a block diagram of a topic operation device for an online content service, shown in accordance with an exemplary embodiment;

fig. 15 is a block diagram illustrating a video topic recommendation device, according to an example embodiment.

Detailed Description

Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the invention. Rather, they are merely examples of apparatus and methods consistent with aspects of the invention as detailed in the accompanying claims.

FIG. 1 is a schematic diagram of an environment in which the present invention may be practiced. In one exemplary embodiment, the implementation environment includes a user terminal 110 implementing an online content service, a content server 130, and a topic operation server 150.

The user terminal 110 provides content to the user through implementation of an in-application online content service. For example, the application is a multimedia application such as a video application, but may be an application for providing content such as text, and is not limited thereto. The user is enabled to obtain respective resources in the internet through the online content service by providing contents in the form of various videos, audios, texts, etc. which are online by the online content service.

The content server 130 interacts with the online content service in data to enable presentation of content for the online content service. For example, for an online content service operated by a video application, the content server 130 is a streaming server.

The topic operation server 150 is used to provide topics related to content to the content server 130, which will be used for the display of topic lists in the topic operation page of the online content service. The topic operation page can be a trending topic list page and the like.

In order to push topics related to content to the online content service, the topic operation server 150 deploys a topic recommendation system, and the topic recommendation system automatically carries out personalized recommendation of topics related to the content to the online content service through topic recommendation in the online content service realized by the method.

Of course, the topic operation server 150 is configured with a manual operation portal in addition to the deployed topic recommendation system, and pushes topics for online content services by manually editing the topics.

It should be appreciated that the content server 130 may be separate from the topic operation server 150 or may be integrated together and may be determined based on the machine deployment requirements of the topic recommendation system.

For users, the deployment of the topic recommendation system can enhance the topic operation efficiency and realize thousands of people and thousands of faces of topic recommendation.

Fig. 2 is a block diagram of an apparatus according to an example embodiment. For example, the apparatus 200 may be the user terminal 110 in the implementation environment shown in fig. 1. For example, the user terminal 110 may be a terminal device such as a smart phone, a tablet computer, a portable computer, a desktop computer, various cameras, and the like.

Referring to fig. 2, the apparatus 200 includes at least the following components: a processing component 202, a memory 204, a power component 206, a multimedia component 208, an audio component 210, a sensor component 214, and a communication component 216.

The processing component 202 generally controls overall operation of the apparatus 200, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations, among others. The processing assembly 202 includes at least one or more processors 218 to execute instructions to perform all or part of the steps of the methods described below. Further, the processing component 202 includes at least one or more modules that facilitate interactions between the processing component 202 and other components. For example, the processing component 202 may include a multimedia module to facilitate interaction between the multimedia component 208 and the processing component 202.

The memory 204 is configured to store various types of data to support operations at the apparatus 200. Examples of such data include instructions for any application or method operating on the apparatus 200. The Memory 204 is implemented by at least any type of volatile or non-volatile Memory device or combination thereof, such as static random access Memory (Static Random Access Memory, SRAM), electrically erasable Programmable Read-Only Memory (Electrically Erasable Programmable Read-Only Memory, EEPROM), erasable Programmable Read-Only Memory (Erasable Programmable Read Only Memory, EPROM), programmable Read-Only Memory (PROM), read-Only Memory (ROM), magnetic Memory, flash Memory, magnetic disk, or optical disk. Also stored in memory 204 are one or more modules configured to be executed by the one or more processors 218 to perform all or part of the steps of any of the methods shown in fig. 3, 4, 5, 6, 7, 8, and 9, described below.

The power supply component 206 provides power to the various components of the device 200. The power components 206 include at least a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 200.

The multimedia component 208 includes a screen between the device 200 and the user that provides an output interface. In some embodiments, the screen may include a liquid crystal display (Liquid Crystal Display, LCD for short) and a touch panel. If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or sliding action, but also the duration and pressure associated with the touch or sliding operation. The screen also includes an organic electroluminescent display (Organic Light Emitting Display, OLED for short).

The audio component 210 is configured to output and/or input audio signals. For example, the audio component 210 includes a Microphone (MIC) configured to receive external audio signals when the device 200 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may be further stored in the memory 204 or transmitted via the communication component 216. In some embodiments, audio component 210 further includes a speaker for outputting audio signals.

The sensor assembly 214 includes one or more sensors for providing status assessment of various aspects of the apparatus 200. For example, the sensor assembly 214 detects the open/closed state of the device 200, the relative positioning of the assemblies, the sensor assembly 214 also detects a change in position of the device 200 or a component of the device 200, and a change in temperature of the device 200. In some embodiments, the sensor assembly 214 further includes a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 216 is configured to facilitate communication between the apparatus 200 and other devices in a wired or wireless manner. The device 200 accesses a WIreless network based on a communication standard, such as WiFi (WIreless-Fidelity). In one exemplary embodiment, the communication component 216 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component 216 further includes a near field communication (Near Field Communication, NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on radio frequency identification (Radio Frequency Identification, RFID) technology, infrared data association (Infrared Data Association, irDA) technology, ultra Wideband (UWB) technology, bluetooth technology, and other technologies.

In an exemplary embodiment, the apparatus 200 is implemented by one or more application specific integrated circuits (Application Specific Integrated Circuit, abbreviated ASIC), digital signal processors, digital signal processing devices, programmable logic devices, field programmable gate arrays, controllers, microcontrollers, microprocessors or other electronic components for executing the methods described below.

Fig. 3 is a flowchart illustrating a topic recommendation method in an online content service according to an exemplary embodiment. In one exemplary embodiment, the topic recommendation method in the online content service, as shown in fig. 3, includes at least the following steps.

In step 310, the topic recommendation system obtains, for a given content of the online content service, base information of the given content and topic information corresponding to the given content.

It should be noted that, first, the online content service is to obtain and present content online through the internet, for example, pull content to a content server through the internet, or receive push content from the content server, so as to obtain and present content. Through an online content service, users are enabled to obtain content that is available for further viewing or browsing. For example, in a video application, by implementing an online content service, a movie video can be obtained in real time, and a user can watch video content by clicking the movie video.

In addition to enabling presentation of content, online content services also perform retrieval and presentation of information related to the content, such as topic information. That is, in the page display performed by the online content service, a topic operation page is included in addition to the display of the content page, so that it is needed to implement topic recommendation for the online content service.

And the topic recommendation system, as described in the foregoing exemplary embodiments, is implemented by the topic operation server. The online content service is realized by interaction with a content server, the content server receives a content request of the online content service, the requested content is given content, and at the moment, the topic recommendation system obtains the given content through interaction with the content server, so that topic recommendation of the given content in the online content service is initiated. The topic operation server of the topic recommendation system can be independently deployed by the content server, namely, the content server can be respectively deployed on different machines, but can also be integrated together, so that the content provision and the topic recommendation can be realized on the same machine.

Topic recommendation by the topic recommendation system for the online content service is directed to a given content, which is referred to herein as content designated for topic recommendation. For example, the given content may be popular content, content of interest to the user, and so forth. In one exemplary embodiment, the given content is in the background of an online content service, i.e., as specified by the content server as referred to by the aforementioned implementation environment.

In one exemplary embodiment, a video server corresponding to an online video content service, as one of the content servers, downloads content designation information that designates a given content to a topic recommendation system for a topic presentation required for the online video content service to be performed, so that the performed topic recommendation can obtain the given content.

Each content is not limited to a given content, and is described by the corresponding basic information. In one exemplary embodiment, for video content, the underlying information includes the director of the video, director information, video point of view information, and the like.

The topic information corresponding to the content refers to information of topic points and content, wherein the topic information is created and shared and transmitted by users in the internet, and as the users participate in the discussion of the topic information, the users further create and share and transmit new topic information. It should be understood that topic information corresponding to content is a discussion topic related to the content and is also generated by the content. Each topic information is shared and propagated with heat.

The topic information corresponding to the given content is acquired by the topic recommendation system through the internet, so that when topic recommendation is required for the given content, the corresponding topic information is only required to be directly acquired according to the given content.

By now it can be appreciated that topic information corresponding to other content is not limited to a given content, but is also obtained and stored with the information crawling conducted towards the internet in order to perform the corresponding topic recommendation process.

In one exemplary embodiment, topic information corresponding to a given content may come from multiple sources such as news, microblogs, social networks, etc., and to ensure topic recommendations that are made over time and correspond to various different content, crawling of content-related topic information is performed for multiple sources that are accessible. The content which can be presented by the online content service is captured and stored with relevant topic information.

In an exemplary embodiment, prior to step 310, the method further comprises: the topic recommendation system captures network information related to content from the internet according to the content distributed by the online content service, the network information related to the content forming basic topic data corresponding to the content, and the basic topic data is used for providing topic information for given content.

The online content service is used for realizing content distribution of users, namely, efficiently distributing content for the users. Thus, online content services are distributed content, i.e., content presented to users. The topic recommendation system performs crawling of content-related network information, such as news, microblog messages, social network messages, etc. related to crawling content, as indicated above, based on content distribution by the online content service. The network information captured for content precipitation forms basic topic data for use by the topic recommendation process that needs to be performed.

In response to this, the topic recommendation system acquires basic information for each content, and also acquires related topic information from the basic topic data for the content, together with topic recommendation for each content.

With the continuous information capturing, the topic information acquired for a given content may be, in an exemplary embodiment, the current latest topic information and the most popular topic information, so as to ensure timeliness and popularity of the subsequent recommended topics.

In step 330, a correlation calculation of topic information with the base information is performed for the given content based on the base information and the topic information, and candidate topic information is obtained from topic information corresponding to the given content based on the calculated correlation.

The topic recommendation system performs topic recommendation on the online content service, and in one exemplary embodiment, the topic recommendation system may be a single content or multiple content, which is not limited herein.

After the basic information and the topic information are obtained for each given content by performing the foregoing step 310 in the topic recommendation system, the topic information may be screened according to the basic information, and several candidate topic information may be obtained for the given content.

As indicated in the foregoing description, topic information is captured from the internet by the topic recommendation system according to the content, and thus, the correlation between the captured topic information and the content is to be determined. For example, the captured topic information is often not related to the content, but only keywords corresponding to the content are introduced into the topic information.

Therefore, it is necessary to determine the correlation between the topic information and the base information, and the level of the correlation between the topic information and the base information indicates the correlation between the topic information and the content. It is necessary to screen out topic information most relevant for given content as candidate topic information according to the correlation between topic information and basic information.

In step 350, the topic recommendation system obtains a recommended topic for the group of users according to the group feature information related to the users and the candidate topic information corresponding to the given content, and distributes the recommended topic for the group of users.

After screening and obtaining a plurality of candidate topic information for each given content, the topic recommendation system can adapt to the user in the obtained plurality of candidate topic information to obtain personalized recommended topics, and further personalized topic recommendation is achieved for the user.

The indicated user is the user logged in by the online content service. Topic recommendations are made for different users. In one exemplary embodiment, the user is presented in the form of user account information.

The method is suitable for basic attributes of users and characteristics of viewing content browsed by the users, and recommended topics are screened from the obtained candidate topic information. In order to adapt to users and realize accurate screening in candidate topic information, on one hand, the basic attributes of the users and the characteristics of the viewing content browsed by the users as screening basis can come from the group where the users are located, namely group characteristic information related to the users, so as to avoid that topic information related to the users can not be screened out due to too little available basis; on the other hand, the comprehensive accuracy of screening is ensured under the control of group characteristic information.

It should be appreciated that the user-related community characteristic information is used to describe the attribute characteristics corresponding to the community of users. The group characteristic information is obtained by classifying the logged-in users of the online content service, and then extracting the basic characteristics of the users under the user group and the use characteristics of the online content service.

Taking an online video content service as an example, all corresponding video users have basic attributes, and along with the control of the online video content service by the video users, video viewing characteristics such as viewing history information, search history, attention records and the like are also reserved, user portrait information is generated for the video users according to the basic attributes and the video viewing characteristics, and then the video users are clustered according to the user portrait information, so that different user groups are obtained.

And extracting group characteristics from the obtained user group to obtain group characteristic information. For example, the partitioned user groups may include a small city male student, a middle aged chasing major, a young office female, and the like.

In the topic recommendation performed for a user, after the plurality of candidate topic information is obtained through the execution of the steps, the user can be subjected to screening of recommended topics in the candidate topic information according to the related group characteristic information, the screened recommended topics are distributed to the user through the online content service, and of course, other users belonging to the same group with the user can also obtain the recommended topics through the online housekeeping service operated by the user.

The topic recommendation system performs recommended topic screening in a plurality of candidate topic information according to the group characteristic information, and extracts candidate topics which are most interested in the group of the user from the candidate topic information as recommended topics. Topic recommendation is realized for users aiming at user groups, and richness of topic recommendation is enhanced while accuracy and effectiveness of recommendation are guaranteed.

According to the method and the device for recommending the content, the personalized recommendation with rich topics can be carried out for each given content, the operation of the online content service is not limited by manual editing of the topics, recommended topics suitable for users can be automatically obtained, and the topic operation performance is enhanced.

Fig. 4 is a flow chart depicting step 310, according to a corresponding embodiment of fig. 3. In one exemplary embodiment, as shown in FIG. 4, this step 310 includes at least:

in step 311, the topic recommendation system reads, from the underlying topic data corresponding to the content, network information associated with the given content, including the given content, based on the given content of the online content service.

In step 313, high-heat topics are extracted from the network information related to the given content, the extracted high-heat topics forming topic information corresponding to the given content.

In the topic recommendation, the content concerned by the user or the content promoted in the background is taken as the given content, and the reading of the network information related to the given content is carried out in the basic topic data of the network information related to the deposited and grabbed content.

For each given content, there is a large amount of network information related thereto, and therefore, it is necessary to extract network information as a high-heat topic point as topic information according to the heat of the corresponding topic point.

With this exemplary embodiment, network information captured by multiple parties in the internet is preprocessed to obtain topic information capable of topic recommendation for a given content.

It should be appreciated that for the progress of a topic recommendation, there are stored a number of underlying topic data, which are all corresponding to a certain content, and that a given content for which a topic recommendation is requested is a part of the content, i.e. all the content corresponding to the stored underlying topic data includes the given content.

Fig. 5 is a flow chart depicting step 330, according to the corresponding embodiment of fig. 3. In one exemplary embodiment, as shown in FIG. 5, this step 330 includes:

in step 331, the topic recommendation system uses the corresponding basic information and topic information as input layer information of the deep neural network for given content, and performs the deep neural network calculation to obtain the probability that each topic information is related to the basic information.

In step 333, candidate topic information is obtained from topic information according to the probability that the topic information is related to the base information.

In the topic recommendation performed for the user, topic information corresponding to the given content and basic information of the given content are obtained through the execution of the step 310, on the basis of the topic information, the topic recommendation system performs correlation calculation between the topic information and the basic information for the given content through the trained deep neural network, and the process is realized through model calculation performed by the deep neural network.

In other words, in the trained deep neural network, for each given content, basic information and topic information are used as input layer information on the input layer, and the probability that each topic information is related to the basic information is obtained through the model calculation, which is the degree of correlation between the basic information and the topic information. By doing so, the decision operation of the deep neural network is performed, and the topic information most relevant to the given content is classified and obtained, and the topic information is used as candidate topic information.

With this exemplary embodiment, the filtering of candidate topic information is achieved by means of a deep neural network, whereby content relevance of the filtered topics is guaranteed for subsequent topic filtering towards the user.

It should be understood that in an exemplary embodiment, the model calculation of the deep neural network is performed on the basic information and the topic information, which is essentially performing vectorization on the basic information and the topic information, and then taking the vector similarity as a content-topic relevance score, where the content-topic relevance score is the output layer information of the deep neural network, so that the output layer information can obtain a plurality of topic information with the highest relevance as candidate topic information.

Fig. 6 is a flow chart depicting step 350, in accordance with the corresponding embodiment of fig. 3. In one exemplary embodiment, this step 350, as shown in FIG. 6, includes at least the following steps.

In step 351, the topic recommendation system obtains group feature information related to users, which are target users for topic recommendation.

In step 353, the deep neural network model calculation is performed on the candidate topic information of the given content according to the group feature information related to the user, so as to obtain the recommended topic of the group in which the user is located.

Among topic recommendations made for users, the referred user is the user who logs in to the online content service, and is also the target user of the topic recommendation system for topic recommendation on the online content service. As described above, in the content presentation performed by the online content service, the corresponding user portrait information is dynamically generated in accordance with the content viewing, content searching, content attention, and the like performed by the user, and the user grouping is continuously performed based on this.

Therefore, in the topic recommendation performed by the topic recommendation system for the user, the user is directed to extracting the recommended topics after obtaining the candidate topic information for the given content by performing the steps 310 to 330.

In this exemplary embodiment, the acquisition of the recommended topics is achieved by deep neural network model calculations. For each given content, performing deep neural network computation on corresponding candidate topic information according to the relevant group characteristic information. That is, the correlation between each candidate topic information and the group feature information, that is, the information output by the output layer, is obtained by the deep neural network model calculation performed by using the group feature information and the candidate topic information related to the user as the input information of the input layer, and the recommended topics are extracted from the plurality of candidate topic information based on the correlation.

It should be understood that, according to the group feature information related to the user, the content focused by the user can be obtained, and the topic information which can be recommended to the user, namely the recommended topic, is obtained by obtaining the candidate topic information corresponding to the content.

The resulting recommended topics may be output to the user through the online content service, such as by presentation of a list of recommended topics in the implemented online content service.

Fig. 7 is a flow chart depicting step 351 according to the corresponding embodiment of fig. 6. In one exemplary embodiment, this step 351 includes at least the following steps.

In step 401, the topic recommendation system obtains a user's underlying features as well as content acquisition features that describe the user's behavioral habits with respect to content distributed by the online content service.

In step 403, user grouping is performed according to the basic features and the content acquisition features, and group feature information corresponding to the group in which the user is located is extracted.

Wherein the user's basic characteristics are used to describe the user's basic conditions, such as gender, age, interests, etc. For users logged in by an online content service, storage of basic features is necessarily performed. For the actions triggered by the user in the online content service, such as content watching actions, content searching actions, content attention actions, etc., there are corresponding records, and the records generate the content acquisition features corresponding to the user.

In the topic recommendation, candidate topic information of a given content needs to be screened by group feature information, and group feature information corresponding to one user group is extracted after grouping users by basic features and content acquisition features. The group characteristic information is the characteristic description of the user corresponding to the group in which the user is located, and on one hand, the group characteristic information describes the basic attribute of the user in the online content service, namely, the basic attribute is matched with the basic characteristic, and on the other hand, the behavior habit of the user in the group triggered by the online content service and related to the content is described. Thus, community characteristic information is strongly correlated not only with users, but also with online content services.

By the above-described exemplary embodiments, the recommendation of topics is realized on a plurality of applications or platforms for content distribution, that is, applications or platforms for deploying online content services, for example, online video content services, so that corresponding topic information can be automatically recommended according to different given contents, manual editing of topics is not required for this purpose, and the topics that can be presented are not limited any more and are limited by human factors.

The invention also provides a topic operation method of the online content service, which is suitable for the topic recommendation method in the online content service. In one exemplary embodiment, as shown in fig. 8, the topic operation method of the online content service at least includes:

in step 510, the topic operation server obtains a content request sent by the online content service, where the content request carries user account information.

Wherein the online content service performs content presentation by front-end display performed at the user terminal. When the online content service is started, a content request is sent to the background to request a content list, at this time, the background returns the content list and basic information of each content in the list, wherein the content is given content, and for the background, the basic information returned to the online content service corresponds to user account information, that is, the background must record the content returned by the background for the user.

Therefore, corresponding to the online content service, the topic operation system implemented by the background deployment can acquire the content request sent by the online content service along with the starting and triggered topic presentation of the online content service, so that the content returned by the background to the online content service can be known according to the user account information carried in the content request, namely the given content.

For example, a video application implementing an online video content service, once started, initiates a content request to the background to obtain video content that can be presented in the started video application, and the background is corresponding to the video content returned by the user, and necessarily records according to user account information.

The user account information is used to indicate user identification information of a user logged in to the online content service, for example, a user account, etc., to uniquely identify the user in the content distribution performed.

In step 530, the topic recommendation system deployed in the topic operation server returns the basic information and topic information of the given content to the online content service according to the content request, performs correlation calculation on the basic information and topic information of the given content, and obtains candidate topic information from topic information according to the calculated correlation.

As described above, given content returned by the background to the online content service logged in by the user is determined according to the user account information carried by the content request, and then basic information and topic information of each given content are obtained according to the given content.

And carrying out correlation calculation on each piece of topic information according to the basic information of each piece of given content, taking the topic information with highest correlation screened out from the topic information as candidate topic information, and the like, so as to finish topic information screening of all pieces of given content, and further obtain a plurality of candidate topic information of each piece of given content.

In step 550, the topic recommendation system obtains, from the candidate topic information, a recommended topic for the group in which the user account information indicates the user, according to the group feature information corresponding to the user account information carried by the content request.

As mentioned above, the user distribution is dynamically performed along with the triggering of the user on the online content service, and then the group characteristic information dynamically adapted to the current user is obtained from the user account information. The screening of topic information of given content based on group characteristic information is necessarily directed to the group in which the user is located.

The topic recommendation is performed for massive users, and the actions of each user on the online content service trigger are dynamically changed at all times, so that the user grouping and the dynamic extraction of group characteristic information which are dynamically changed at all times are needed, and the topic recommendation for the group of the user can be realized at the moment by performing the topic recommendation for the user, which can make the topic recommendation simple and quick, and save the computing resources.

In step 570, the topic operation server merges the recommended topics and the manually operated topics of the given content to obtain topic data, and returns the topic data to the online content service via the content server according to the user account information.

Firstly, it should be noted that the manual topic operation performed by the topic operation server not only performs topic recommendation automatically through the deployed topic recommendation system, but also performs manual topic operation, namely, manually editing topics, so as to provide manual intervention for topic presentation of users.

Based on the method, the topic operation server carries out automatic topic recommendation through the deployed topic recommendation system to obtain recommended topics, and on the other hand, the topic operation server also obtains manual operation topics. For topic recommendation performed for a user, the recommended topics and the manually operated topics form topic data for presenting topics to the user, so that the topic data is returned to the online content service through the content server according to the user account information.

The fusion of the recommended topics and the manual operation topics in the topic operation server is performed according to the product strategy of the online content service, for example, the manual operation topics are returned preferentially, and the recommended topics are returned subsequently, so that the manual operation topics are displayed preferentially through the head part of the topic list displayed by the online content service.

In addition, a certain intervention configuration can be set, and manual intervention is performed on the obtained topic data, for example, when the recommended topics have low relevance to the topics due to deviation, or sensitive contents appear on the recommended topics, manual intervention is needed to be removed.

In another exemplary embodiment, before step 530, the topic operation method of the online content service further includes:

and the topic recommendation system in the topic operation server acquires basic information and topic information of given content returned to the online content service according to the user account information carried in the content request, wherein the topic information corresponds to the given content.

In the process of topic operation, if topics need to be recommended to users, firstly acquiring a content request initiated after the online content service logged in by the users is started, acquiring given content returned to the users according to the content request, and further acquiring basic information and topic information corresponding to the given content.

In another exemplary embodiment, before the step of the topic operation server obtaining basic information and topic information of the given content returned to the online content service according to the user account information carried in the content request, the topic operation method of the online content service further includes:

the topic recommendation system captures network information related to content in the internet according to the content distributed by the online content service, the network information related to the content forms basic topic data corresponding to the content, and the basic topic data is used for providing topic information for given content.

In this case, as described above, the user-oriented content distribution is realized by the online content service, and the topic recommendation to be performed is performed, so that the network information related to the content is captured according to the content to be distributed, and basic topic data corresponding to the content is formed, and the basic topic data is used for providing topic information.

Through the above-mentioned exemplary embodiments, a topic operation system is implemented for online content service, and the topic operation system includes a topic recommendation system, so that topic presentation in online content service is implemented through docking of the topic operation system and a content server, and efficient and accurate topic recommendation is continuously implemented for users.

Correspondingly, the invention also provides a video topic recommendation method for recommending video topics for the online video content service. Fig. 9 is a flowchart illustrating a video topic recommendation method according to an example embodiment. In an exemplary embodiment, as shown in fig. 9, the video topic recommendation method at least includes:

in step 710, the video topic recommendation system obtains video base information for a given video content and topic information corresponding to the given video content for the given video content of the online video content service.

It should be noted that, first, the online video content service is one of the online content services, and the corresponding given video content is one of the given content. The online video content service obtains video content online through a video server, and the online video content server obtains a recommended topic returned for the given video content from a video topic recommendation system through the video server. As described above, the video topic recommendation system will be deployed in a topic operation server that will push corresponding video topics for a given video content returned by the video server to an online video content service. The pushed video topics include recommended topics obtained by the video topic recommendation system, and other related video topics can be included in addition to the recommended topics, which are not limited herein.

In video topic recommendation by a video topic recommendation system for an online video content service, first, video base information and video topics, that is, topic information corresponding to a given video content, are acquired for the given video content.

In one exemplary embodiment, the acquired video base information includes a video title, a video ID, a video picture, a date of putting on shelf, and the like to describe the corresponding video content.

In step 730, a correlation calculation of topic information and base information is performed for a given video content based on the video base information and topic information, and candidate topic information is obtained from topic information corresponding to the given video content based on the calculated correlation.

Corresponding to the candidate topic screening performed in the foregoing exemplary embodiment, each topic information is screened for each given video content according to the video basic information, where the screening is implemented by correlation calculation, so as to screen the topic information most relevant to the video basic information as the candidate topic information.

In step 750, according to the group feature information related to the video user and the candidate topic information corresponding to the given video content, a recommended topic for the group of the video user is obtained for the given video content, and the recommended topic is distributed for the group of the video user.

According to the video topic recommendation system, the video topic is recommended online for the online video content service realized by the video application, so that the video topics suitable for users can be accurately and timely pushed to the user terminal, the need of waiting for manual editing of the topics is avoided, and the operation performance of the video topics is enhanced.

In another exemplary embodiment, the video topic recommendation method further includes, prior to step 710: the video topic recommendation system selects topic presentation to be performed according to the online video content service and receives a content request sent by the online video content service;

wherein the content request is for a given video content for which a logged-in video user for an online video content service indicates a recommendation of a video topic.

That is, as the presentation of video content in an online video content service proceeds, a user may view the presented video content and initiate a video topic recommendation via a content request with this video content as given video content.

Therefore, the video topic recommendation is performed towards the video content watched by the user, the video topic presentation and discussion are performed around the video content watched by the user at present, even the video content watched by the user, and then new video topics are continuously induced.

Through the above-mentioned exemplary embodiments, the recommendation of video topics is realized for the video application, in particular for the online video content service operated by the video application, and then before the basis, the operation of the video topics can be implemented by fusing the manual topic operation as realized by the topic operation system, so that the existence of manual intervention is avoided.

Taking an online video content service as an example, the method is combined to realize video topic operation, so that online vision is realized The video content service presents video topics on the user terminal.

In a video APP (Application) operated by a user terminal, video content is requested to the background through the start of an online video content service, at this time, a content server returns the video content to the user terminal on one hand, and on the other hand, a topic recommendation system is triggered through the content request to realize the video topic display on the user terminal.

Fig. 10 is a system architecture diagram of a video topic operation system shown in accordance with an exemplary embodiment. In one exemplary embodiment, as shown in FIG. 10, a content request initiated by a video APP to the background will be sent to the topic logic service 910 running in the background.

The video topic operation system includes a topic logic service 910, a topic recommendation system 930, a manual operation topic storage service 950, and a topic operation service 970.

The video APP is used for providing video services, namely online video content services as indicated above, for video information display, video playing, topic display and user interaction to video users.

The topic logic service 910 is configured to receive a topic list request of a video APP, the topic list request being initiated by a background for a given content in response to a content request of the video APP to request topic information related to the given content.

The topic operation service 970 is oriented to operators to obtain manual operation topics under manual collection of operators, and stores the manual operation topics through the manual operation topic storage service 950 to form a basic topic data pool.

The topic recommendation system 930 generates a basic recommended topic list, such as a hot topic list related to video content, based on basic features of the user and basic information of the video content by topic recommendation based on the deep neural network.

While for topic recommendation system 930, intelligent topic recommendation is implemented through a deep neural network in accordance with an exemplary embodiment of the present invention. Fig. 11 is a system architecture diagram of the topic recommendation system shown in accordance with the corresponding embodiment of fig. 10.

As shown in fig. 11, the topic recommendation system 930 mainly includes a topic crawling service 931, a candidate topic screening service 933, a user feature grouping service 935, and a recommended topic ranking service 937;

in addition, the topic recommendation system 930 further includes a hot topic information storage service and a video information storage service to support the candidate topic screening service 933, a search history storage service 1005 to support the user feature grouping service 935, a viewing history storage service 1007, a user base feature storage 1009, and the like.

Optionally, the topic crawling service 931 is configured to crawl trending topics, such as news portals, microblog headline messages, and friend circle messages, from the internet for a plurality of video content, respectively.

The hot topic information storage service 1001 is configured to store topic information obtained by grabbing by the topic grabbing service 931, and form a basic topic data pool.

The video information storage service 1003 is used to store basic information related to video, for example, a director of video content, director information, point of view information of video content, and the like.

The candidate topic screening service 933 calculates basic information of the video content and the captured topic information through the deep neural network to obtain candidate topic information of topN generated corresponding to the video content.

The user basic feature storage service 1009 is used to store basic features of the user, such as gender, age, region, occupation, and the like. The viewing history storage service 1007 is used to store history information of viewing video content by a user. The search storage service 1005 is used to store a search record of the user on the video APP.

The user feature clustering service 935 generates portraits for users based on information such as the user's basic features, viewing history information, search history, attention records, etc., and clusters users based on user images to extract group feature information.

The recommended topic ranking service 937 is also implemented by a deep neural network, and the input layer information of the recommended topic ranking service is candidate topic information related to video content and group feature information related to users, so as to screen topics meeting the interests of the users from the candidate topic information according to the features of the users, so as to obtain a topic information list finally recommended to the users.

Fig. 12 is a timing diagram for video topic generation according to the corresponding embodiment of fig. 11. That is, as shown in fig. 12, the video APP initiates a topic list request to the topic logic service by initiating a content request of the video APP, that is, as shown in step 1, the request carries user account information and corresponding video id information.

Video id information for indicating given video content, i.e., video content returned to the video APP in response to a content request by the background.

As indicated in step 2, the topic logic service initiates an asynchronous request after receiving the request. The topic logic service requests the manual operation topic storage service to acquire manual intervention information and a topic list of manual operation.

On the other hand, under the execution of step 3, the topic logic service also requests the recommended topic ranking service, from which automatic generation of recommended topics is requested.

At this time, after the manual operation storage service receives the request, the manual intervention information and the topic list information of the manual operation are returned, which is shown in step 4.

After the recommended topic sorting service receives the request, in step 5, an asynchronous request is initiated, and group feature information related to the user is requested to the user feature grouping service according to the user account information.

The recommended topic sorting service executes step 6, requests candidate topic list information from the candidate topic screening service according to the basic information related to the video content, and after receiving the request, the candidate topic screening service executes step 7, and requests the basic information of the video content from the video information storage service according to the content identification information.

The candidate topic screening service executes step 8 at the same time, initiates a request to the hot topic storage service, and reads hot topic information; after receiving the recommended topic ranking service request, the user feature grouping service returns group feature information corresponding to the user in real time according to an interface of offline training, namely, in step 9.

Step 10, after receiving the request of the candidate topic screening service, the video information storage service returns the basic information corresponding to the video in real time according to the video id, and then step 11 is executed, namely, after receiving the request of the candidate topic screening service, the hot topic storage service returns the stored hot topic list in real time.

After the candidate topic screening service receives the video basic information and the hot topic list, the candidate topic screening service performs characterization processing, and then inputs deep neural network training, as in step 12.

As described in steps 13 to 15, the candidate topic screening service returns a candidate topic list according to the deep neural network training result; after receiving the user characteristic information and the candidate topic list service, the recommended topic ordering service inputs a deep neural network for training; the recommended topic ranking service returns a recommended topic list according to the deep neural network training result.

Finally, under the execution of step 16, the topic logic service acquires the manual intervention information, the manual operation topic list and the recommended topic list, performs topic list fusion and manual intervention logic processing, and returns a final topic list to the video APP.

Through the implementation process, comments, barrages, news and the like related to video content in the internet are grabbed, topic points with higher heat are extracted, meanwhile, user portraits are created by means of video watching features of video users, and interested video topics are generated for the users by means of model training of the deep neural network according to basic features of the users, basic information of the videos and topic information.

For the same video content, the focused video discussion topics are accurately pushed for different user groups, and the activity of users is improved to the greatest extent.

Through automatic, efficient and accurate video topic operation in online video content service, users are effectively attracted to increase the use time of video application, and video topics are accurately pushed.

The following is an embodiment of the apparatus of the present invention, which is used to execute the topic recommendation method embodiment in the online content service of the present invention. For details not disclosed in the embodiment of the apparatus of the present invention, please refer to the topic recommendation embodiment in the online content service of the present invention.

Fig. 13 is a block diagram illustrating a topic recommendation device in an online content service according to an exemplary embodiment. In one exemplary embodiment, as shown in fig. 10, topic recommendation devices in the online content service are configured to user terminals, and the video recognition devices include, but are not limited to: an information acquisition module 1110, a candidate acquisition module 1130, and a recommendation generation module 1150.

An information acquisition module 1110 for acquiring, for a given content of an online content service, base information of the given content and topic information corresponding to the given content;

a candidate filtering module 1130, configured to perform, for the given content, a correlation calculation between the topic information and the base information according to the base information and the topic information, and obtain candidate topic information from topic information corresponding to the given content according to the calculated correlation;

the recommendation generation module 1150 is configured to obtain a recommendation topic oriented to a group where a user is located according to group feature information related to the user and candidate topic information corresponding to the given content, and distribute the recommendation topic oriented to the group where the user is located.

Fig. 14 is a block diagram of a topic operation device of an online content service according to an exemplary embodiment. In one exemplary embodiment, as shown in fig. 14, the topic operation device of the online content service at least includes: the request acquisition module 1210, the candidate computation module 1230 recommends acquisition module 1250 and the fusion module 1270.

A request acquisition module 1210, configured to acquire a content request sent by an online content service, where the content request carries user account information;

a candidate calculating module 1230 for returning basic information and topic information of a given content to the online content service according to the content request, performing correlation calculation on the basic information and topic information of the given content, and obtaining candidate topic information from the topic information according to the calculated correlation;

the recommendation obtaining module 1250 is configured to obtain, from the candidate topic information, a recommendation topic that indicates a group in which a user is located for the user account information according to group feature information corresponding to the user account information carried by the content request;

the fusion module 1270 is configured to fuse the recommended topics and the manual operation topics of the given content to obtain topic data, and return the topic data to the online content service according to the user account information.

Fig. 15 is a block diagram illustrating a video topic recommendation device, according to an example embodiment. In one exemplary embodiment, as shown in fig. 15, the video topic recommendation device at least includes: a video related information acquisition module 1310, a candidate topic acquisition module 1330, and a recommended topic acquisition module 1350.

A video related information obtaining module 1310, configured to obtain, for a given video content of an online video content service, video base information of the given video content and topic information corresponding to the given video content;

the video topic screening module 1330 is configured to perform, for the given video content, a correlation calculation between the topic information and the base information according to the video base information and topic information, and obtain candidate topic information from topic information corresponding to the given video content according to the calculated correlation;

the video topic secondary screening module 1350 is configured to obtain a recommended topic for a group where a video user is located according to group feature information related to the video user and a plurality of candidate topic information corresponding to the given video content, and distribute the recommended topic for the group where the video user is located.

Optionally, the present invention further provides an electronic device, which may be used in the implementation environment shown in fig. 1, to perform all or part of the steps of the method shown in any of fig. 3, fig. 4, fig. 5, fig. 6, fig. 7, fig. 8, and fig. 9. The device comprises:

a processor;

A memory for storing processor-executable instructions;

wherein the processor is configured to perform a method implementing the previously indicated.

The specific manner in which the processor of the apparatus in this embodiment performs the operations has been described in detail in relation to the previous embodiments and will not be described in detail here.

In an exemplary embodiment, a storage medium is also provided, which is a computer-readable storage medium, such as may be a transitory and non-transitory computer-readable storage medium including instructions. Such as memory 204 including instructions executable by processor 218 of apparatus 200 to perform the methods described above.

It is to be understood that the invention is not limited to the precise arrangements and instrumentalities shown in the drawings, which have been described above, and that various modifications and changes may be effected without departing from the scope thereof. The scope of the invention is limited only by the appended claims.

Claims

1. A method of topic recommendation in an online content service, the method comprising:

for given content of an online content service, basic information of the given content and topic information corresponding to the given content are acquired, topic information related to the given content is captured by an internet oriented access multiparty source, and the given content is subjected to attribute description through the corresponding basic information;

2. The method of claim 1, wherein the given content of the online content service, before obtaining the base information of the given content and the topic information corresponding to the given content, further comprises:

capturing network information related to the content from an internet according to the content distributed by the online content service, wherein the network information related to the content forms basic topic data corresponding to the content, and the basic topic data is used for providing the topic information for the given content.

3. The method of claim 2, wherein the obtaining, for a given content of the online content service, base information of the given content and topic information corresponding to the given content comprises:

Reading network information related to the given content from basic topic data corresponding to the given content according to the given content of the online content service, wherein the content comprises the given content;

extracting high-heat topics from the network information related to the given content, wherein the extracted high-heat topics form topic information corresponding to the given content.

4. The method according to claim 1, wherein the performing, for the given content, a correlation calculation of the topic information with the base information based on the base information and topic information, and obtaining candidate topic information from topic information corresponding to the given content based on the calculated correlation, includes:

for the given content, taking the corresponding basic information and topic information as input layer information of a deep neural network, and executing the deep neural network model calculation to obtain the probability that each topic information is related to the basic information;

and obtaining the candidate topic information from the topic information according to the probability that the topic information is related to the basic information.

5. The method according to claim 1, wherein the obtaining the recommended topics for the group of users according to the group feature information related to the users and the candidate topic information corresponding to the given content includes:

Acquiring group characteristic information related to users, wherein the users are target users for topic recommendation;

and carrying out deep neural network model calculation on the candidate topic information of the given content according to the group characteristic information related to the user, and obtaining recommended topics of the group where the user is located.

6. The method of claim 5, wherein the obtaining user-related group feature information comprises:

acquiring basic characteristics and content acquisition characteristics of the user, wherein the content acquisition characteristics describe behavior habits of the user related to content distributed by the online content service;

and carrying out user grouping according to the basic characteristics and the content acquisition characteristics, and extracting group characteristic information corresponding to the group in which the user is located.

7. A topic operation method of an online content service, the method comprising:

returning basic information and topic information of given content to the online content service according to the content request, performing correlation calculation on the basic information and topic information of the given content, and obtaining candidate topic information from the topic information according to the calculated correlation, wherein the topic information related to the given content is captured by an internet facing to an accessed multiparty source, and the given content is subjected to attribute description through the corresponding basic information;

8. The method of claim 7, wherein the returning of the base information and topic information of a given content to the online content service according to the content request and performing a correlation calculation on the base information and topic information of the given content, the method further comprises, before obtaining candidate topic information from the topic information according to the calculated correlation:

and obtaining basic information and topic information of given content returned to the online content service according to the user account information carried in the content request, wherein the topic information corresponds to the given content.

9. The method of claim 8, wherein before obtaining the base information and topic information of the given content returned to the online content service according to the user account information carried in the content request, the method further comprises:

Capturing network information related to content in an internet according to the content distributed by the online content service, wherein the network information related to the content forms basic topic data corresponding to the content, and the basic topic data is used for providing the topic information for the given content.

10. A video topic recommendation method, the method comprising:

for given video content of an online video content service, acquiring video basic information of the given video content and topic information corresponding to the given video content, wherein the topic information related to the given content is captured by an internet-oriented access multiparty source, and the given content is subjected to attribute description through the corresponding basic information;

11. The method of claim 10, wherein prior to obtaining video base information for a given video content of an online video content service and topic information corresponding to the given video content, the method further comprises:

selecting topic presentation to be performed according to the online video content service, and acquiring a content request sent by the online video content service;

wherein the content request is for a given video content indicating a video topic recommendation for a video user logged in to the online video content service.

12. A topic recommendation device in an online content service, the device comprising:

the information acquisition module is used for acquiring basic information of given content and topic information corresponding to the given content for given content of an online content service, wherein the topic information related to the given content is captured by an internet-oriented access multiparty source, and the given content is subjected to attribute description through the corresponding basic information;

13. A topic operation apparatus for an online content service, the apparatus comprising:

the candidate calculation module is used for returning basic information and topic information of given content to the online content service according to the content request, performing correlation calculation on the basic information and topic information of the given content, obtaining candidate topic information from the topic information according to the calculated correlation, capturing the topic information related to the given content by the internet towards an accessed multiparty source, and carrying out attribute description on the given content by the corresponding basic information;

the recommendation acquisition execution module is used for acquiring recommendation topics which are marked by the group where the user is located and are oriented to the user account information from the candidate topic information according to group characteristic information corresponding to the user account information carried by the content request;

14. A video topic recommendation device, the device comprising:

the system comprises a video related information acquisition module, a video content service module and a content management module, wherein the video related information acquisition module is used for acquiring video basic information of given video content and topic information corresponding to the given video content for given video content of an online video content service, the topic information related to the given content is captured by an internet for access to multiparty sources, and the given content is subjected to attribute description through the corresponding basic information;

the recommendation topic obtaining module is used for obtaining recommendation topics oriented to the group where the video user is located according to group characteristic information related to the video user and candidate topic information corresponding to the given video content, and distributing the recommendation topics oriented to the group where the video user is located.

15. A machine apparatus, comprising:

a processor; and

a memory having stored thereon computer readable instructions which, when executed by the processor, implement the method according to any of claims 1 to 11.