CN111708901A

CN111708901A - Multimedia resource recommendation method and device, electronic equipment and storage medium

Info

Publication number: CN111708901A
Application number: CN202010568998.0A
Authority: CN
Inventors: 孙千柱
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2020-06-19
Filing date: 2020-06-19
Publication date: 2020-09-25
Anticipated expiration: 2040-06-19
Also published as: CN111708901B

Abstract

The application relates to the technical field of content push, and discloses a multimedia resource recommendation method, a multimedia resource recommendation device, electronic equipment and a storage medium, which can more accurately locate the current interest point of a user and improve the pertinence and effectiveness of recommended content. The method comprises the following steps: acquiring bullet screen information sent by a user; determining a comment object aimed at by the bullet screen information and the love degree of the user on the comment object based on the bullet screen information; determining interest tags of the users based on the comment objects and the love degrees; and recommending the multimedia resources matched with the interest tags of the users to the users.

Description

Multimedia resource recommendation method and device, electronic equipment and storage medium

Technical Field

The present application relates to the field of content push technologies, and in particular, to a multimedia resource recommendation method and apparatus, an electronic device, and a storage medium.

Background

With the development of network technology, people are increasingly communicating and acquiring various multimedia resources through network applications, such as browsing news, videos, novels and the like through the network. To facilitate the user's viewing, the web application may actively recommend multimedia assets to the user. The existing multimedia resource recommendation method is usually used for recommending according to the release time of the multimedia resource, the current heat of the multimedia resource or the historical browsing record of a user, so that the recommended content is not the content which is interested by the user, and the recommendation effectiveness is low.

Disclosure of Invention

The embodiment of the application provides a multimedia resource recommendation method and device, an electronic device and a storage medium, which can more accurately locate the current interest point of a user and improve the pertinence and effectiveness of recommended content.

In one aspect, an embodiment of the present application provides a multimedia resource recommendation method, including:

acquiring bullet screen information sent by a user;

determining a comment object for which the bullet screen information is directed and the user's likeability for the comment object based on the bullet screen information;

determining interest tags of the user based on the comment objects and the likeness;

and recommending the multimedia resources matched with the interest tags of the users to the users.

receiving bullet screen information sent by a user;

displaying a multimedia resource recommendation list determined based on the barrage information, wherein multimedia resources included in the multimedia resource recommendation list are matched according to the interest tags of the users, and the interest tags of the users are determined based on the comment objects targeted by the barrage information and the love degrees of the users to the comment objects.

sending a resource acquisition request to a server, wherein the resource acquisition request comprises a user identifier, so that the server returns a corresponding multimedia resource recommendation list, the multimedia resource recommendation list comprises multimedia resources matched with interest tags corresponding to the user identifier, and the interest tags are determined based on comment objects, to which bullet screen information sent by a user corresponding to the user identifier is directed, and the love degrees of the comment objects by the user;

and receiving and displaying the multimedia resource recommendation list returned by the server.

In one aspect, an embodiment of the present application provides a multimedia resource recommendation device, including:

the acquisition module is used for acquiring bullet screen information sent by a user;

the information extraction module is used for determining a comment object aimed at by the bullet screen information and the love degree of the user on the comment object based on the bullet screen information;

an interest determination module, configured to determine an interest tag of the user based on the comment object and the likeness;

and the recommending module is used for recommending the multimedia resources matched with the interest tags of the users to the users.

Optionally, the information extraction module is specifically configured to determine the comment object to which the barrage information is directed by at least one of:

extracting a comment object for which the bullet screen information aims from the bullet screen information; or,

and determining the target multimedia resource associated with the bullet screen information as a comment object aimed at by the bullet screen information.

Optionally, when the comment object is extracted from the barrage information, the information extraction module is specifically configured to determine the user's likeability for the comment object by:

extracting emotion keywords from the bullet screen information;

and determining the love degree of the user to the comment object based on the extracted emotional degree key words.

Optionally, when the comment object is determined according to the target multimedia resource associated with the barrage information, the information extraction module is specifically configured to determine the user's likeability for the comment object by:

aiming at each bullet screen information which is sent by the user and is associated with the target multimedia resource, determining the emotion level corresponding to each bullet screen information based on the emotion degree key words contained in each bullet screen information;

and determining the user's preference degree for the target multimedia resource based on the emotion level corresponding to each bullet screen information.

Optionally, the information extraction module is further configured to determine a bullet screen participation degree when the user watches the target multimedia resource based on the number of bullet screen information sent by the user and associated with the target multimedia resource;

the information extraction module is specifically configured to determine the user's preference for the target multimedia resource based on the emotion level corresponding to each bullet screen information and the bullet screen participation.

Optionally, the information extraction module is specifically configured to:

determining the bullet screen sending frequency when the user watches the target multimedia resource based on the quantity of bullet screen information which is sent by the user and is related to the target multimedia resource;

and determining the bullet screen participation degree of the user when watching the target multimedia resource based on the bullet screen sending frequency.

the first receiving module is used for receiving barrage information sent by a user;

and the display module is used for displaying a multimedia resource recommendation list determined based on the barrage information, the multimedia resources included in the multimedia resource recommendation list are matched according to the interest tags of the user, and the interest tags of the user are determined based on the comment objects targeted by the barrage information and the love degrees of the user on the comment objects.

the system comprises a sending module, a receiving module and a processing module, wherein the sending module is used for sending a resource obtaining request to a server, the resource obtaining request comprises a user identifier so as to enable the server to return a corresponding multimedia resource recommendation list, the multimedia resource recommendation list comprises multimedia resources matched with interest tags corresponding to the user identifier, and the interest tags are determined based on comment objects corresponding to bullet screen information sent by users corresponding to the user identifier and the love degrees of the comment objects by the users;

and the second receiving module is used for receiving and displaying the multimedia resource recommendation list returned by the server.

In one aspect, an embodiment of the present application provides an electronic device, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of any one of the methods when executing the computer program.

In one aspect, an embodiment of the present application provides a computer-readable storage medium having stored thereon computer program instructions, which, when executed by a processor, implement the steps of any of the above-described methods.

In one aspect, an embodiment of the present application provides a computer program product or a 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 the processor executes the computer instructions to cause the computer device to perform the method provided in the various alternative implementations described above.

According to the multimedia resource recommendation method, the multimedia resource recommendation device, the electronic equipment and the storage medium, the comment object aimed at by the bullet screen information is obtained by analyzing the bullet screen information sent by the user, the interest tag of the user is determined based on the comment object aimed at by the bullet screen information, and the multimedia resource matched with the interest tag of the user is recommended to the user. Based on the bullet screen information sent by the user, the object which the user is interested in is determined, the current interest point of the user can be more accurately positioned, the recommended multimedia resources are closer to the content which the user really is interested in, the pertinence and the effectiveness of the recommended content are improved, the use viscosity of the user is further improved, and the multimedia resource playing amount and the daily life of the user are also favorably improved.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings needed to be used in the embodiments of the present application will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present application, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

Fig. 1 is a schematic view of an application scenario of a multimedia resource recommendation method according to an embodiment of the present application;

fig. 2 is a flowchart illustrating a multimedia resource recommendation method according to an embodiment of the present application;

fig. 3 is a schematic flowchart of determining an interest tag based on bullet screen information according to an embodiment of the present application;

fig. 4 is a flowchart illustrating a multimedia resource recommendation method according to an embodiment of the present application;

FIG. 5 is a schematic view of an interface for recommending multimedia resources during a process of a user watching a video according to an embodiment of the present application;

fig. 6 is a flowchart illustrating a multimedia resource recommendation method according to an embodiment of the present application;

fig. 7 is a schematic structural diagram of a multimedia resource recommendation device according to an embodiment of the present application;

fig. 8 is a schematic structural diagram of a multimedia resource recommendation device according to an embodiment of the present application;

fig. 9 is a schematic structural diagram of a multimedia resource recommendation device according to an embodiment of the present application;

fig. 10 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application.

For convenience of understanding, terms referred to in the embodiments of the present application are explained below:

multimedia resources: it refers to the integration of multiple media, generally including multiple media forms such as text, sound and image. The multimedia resources in the embodiments of the present application include, but are not limited to, videos, audios (e.g., radio dramas, music, etc.), electronic books, and the like.

Barrage (barrage): refers to a commentary subtitle that pops up when a multimedia asset is viewed over a network. In the process of watching multimedia resources, a user can watch the barrage published by other users and can also send the barrage, and the sent barrage can also be seen by other users, so that interaction is realized in the process of watching multimedia resources.

Interest tag: tags for identifying content of interest to the user, such as the type of movie the user likes, the actors the user likes, etc. In practical application, the interest tag of the user can be obtained by analyzing the behavior data of the user and mining, wherein the behavior data of the user comprises: the behavior data shared and read by the user includes text information, picture information, video information and the like, such as shared information issued by the user, shared information forwarded by the user, and shared information clicked and read by the user. The interest tags can also be set by the user, preset candidate interest tags can be provided for the user to select, the preset candidate interest tags can be set by the user according to the category of the network application and the content of the information to be recommended, if the category of the network application is social, and the information to be recommended is user data, social candidate interest tags can be provided, such as gender, hobby, region and the like, and the pertinence and the effectiveness of the interest tags can be improved through the preset candidate interest tags.

A named entity generally refers to an entity with a specific meaning or strong reference in the text, and generally includes a name of a person, a name of a place, a name of an organization, a date and time, a proper noun, and the like. The concept of named entities can be very broad, and any special piece of text that is needed by a business can be called a named entity.

Named Entity Recognition (NER), a basic task of natural language processing, aims to extract Named Entities from unstructured input text. The discriminant Model CRF is the current mainstream Model of the NER, and its objective function not only considers the input state feature function, but also includes the label transfer feature function.

Deep learning: the concept of deep learning is derived from the research of an artificial neural network, and a multi-layer perceptron with multiple hidden layers is a deep learning structure. Deep learning forms a more abstract class or feature of high-level representation properties by combining low-level features to discover a distributed feature representation of the data. Deep learning is a new field in machine learning research, and the motivation is to establish and simulate a neural network for analyzing and learning of human brain, and to interpret data such as images, sounds, texts and the like by simulating the mechanism of human brain. Common deep learning models include: convolutional Neural Networks (CNN), cyclic Neural networks (RNN), Long Short-Term Memory networks (LSTM), Deep Neural Networks (DNN), Deep Belief Networks (DBNs), and the like. Data propagates in a neural network in two ways, one along the path from input to output, called forward propagation (backward propagation), and the other from output back to input, called backward propagation (backward propagation). In the forward propagation process, input information is processed layer by layer through a neural network and transmitted to an output layer, errors between output values and expectations are described through a loss function, backward propagation is carried out, partial derivatives of the loss function to the weight of each neuron are calculated layer by layer, weight gradient data of the loss function to weight vectors are formed and serve as the basis for updating weight parameters, and training of the neural network is completed in the process of continuously updating the weight parameters.

A Client (Client), also called Client, refers to a program corresponding to a server and providing local services to clients. Except for some application programs which only run locally, the application programs are generally installed on common clients and need to be operated together with a server. After the internet has developed, the more common clients include web browsers used on the world wide web, email clients for receiving and sending emails, and client software for instant messaging. For this kind of application, a corresponding server and a corresponding service program are required in the network to provide corresponding services, such as database services, e-mail services, etc., so that a specific communication connection needs to be established between the client and the server to ensure the normal operation of the application program.

Cloud technology refers to a hosting technology for unifying serial resources such as hardware, software, network and the like in a wide area network or a local area network to realize calculation, storage, processing and sharing of data.

Cloud technology (Cloud technology) is based on a general term of network technology, information technology, integration technology, management platform technology, application technology and the like applied in a Cloud computing business model, can form a resource pool, is used as required, and is flexible and convenient. Cloud computing technology will become an important support. Background services of the technical network system require a large amount of computing and storage resources, such as video websites, picture-like websites and more web portals. With the high development and application of the internet industry, each article may have its own identification mark and needs to be transmitted to a background system for logic processing, data in different levels are processed separately, and various industrial data need strong system background support and can only be realized through cloud computing.

Big data (Big data) refers to a data set which cannot be captured, managed and processed by a conventional software tool within a certain time range, and is a massive, high-growth-rate and diversified information asset which can have stronger decision-making power, insight discovery power and flow optimization capability only by a new processing mode. With the advent of the cloud era, big data has attracted more and more attention, and the big data needs special technology to effectively process a large amount of data within a tolerance elapsed time. The method is suitable for the technology of big data, and comprises a large-scale parallel processing database, data mining, a distributed file system, a distributed database, a cloud computing platform, the Internet and an extensible storage system.

In the application, the storage and management of the multimedia resources can be realized and the multimedia service can be provided for the user based on the cloud technology and the big data technology.

Any number of elements in the drawings are by way of example and not by way of limitation, and any nomenclature is used solely for differentiation and not by way of limitation.

In a specific practical process, the existing multimedia resource recommendation method is often recommended according to the release time of the multimedia resource, the current popularity of the multimedia resource or the historical browsing record of the user, so that the recommended content is often not the content which the user is interested in, and the recommendation effectiveness is very low.

Therefore, the multimedia resource recommendation method provided by the application obtains the comment object aimed at by the bullet screen information by analyzing the bullet screen information sent by the user, determines the interest tag of the user based on the comment object aimed at by the bullet screen information, and recommends the multimedia resource matched with the interest tag of the user to the user. The method and the device have the advantages that the objects which are interested by the user are determined based on the barrage information sent by the user in real time, and the current interest points of the user can be more accurately positioned, so that the recommended multimedia resources are closer to the content which is really interested by the user, the pertinence and the effectiveness of the recommended content are improved, the use viscosity of the user is further improved, and the multimedia resource playing amount and the daily life of the user are also favorably improved.

After introducing the design concept of the embodiment of the present application, some simple descriptions are provided below for application scenarios to which the technical solution of the embodiment of the present application can be applied, and it should be noted that the application scenarios described below are only used for describing the embodiment of the present application and are not limited. In specific implementation, the technical scheme provided by the embodiment of the application can be flexibly applied according to actual needs.

Referring to fig. 1, a schematic view of an application scenario of the multimedia resource recommendation method according to the embodiment of the present application is shown. In the application scenario shown in fig. 1, a terminal device 101 and a server 102 are included. The terminal device 101 and the server 102 may be connected through a wireless communication network or a wired communication network, and the terminal device 101 includes, but is not limited to, a desktop computer, a mobile phone, a mobile computer, a tablet computer, a media player, a smart wearable device, a smart television, a vehicle-mounted device, a Personal Digital Assistant (PDA), and other electronic devices. The server 102 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 basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a network service, cloud communication, middleware service, a domain name service, a security service, a CDN, a big data and artificial intelligence platform, and the like.

A client is installed in the terminal apparatus 101, and the client provides a multimedia service by the server 102. The client may be a multimedia resource client such as a browser client, a video client, an audio client, and a reading client, where the client is served by the server 102, for example, a youku client, an internet music client, and an online reading client, and may also be a player or a reader capable of providing a playing service without networking, and the client may be used to play the multimedia resource stored in the storage space of the terminal device 101, or may be a player or a reader having an independent playing function and capable of using the multimedia resource service provided by the server 102, for example, thousands of mutes, storm videos, and so on, and this is not specifically limited in this embodiment of the present application.

The server 102 is used for providing multimedia services, which may refer to video services, audio services, picture services, reading services, question and answer services, and the like, and multimedia resources include, but are not limited to, video, audio, text, pictures, and the like. Taking a video server as an example, the video service provided by the server may include services such as live video, online video playing, video downloading, and the like, and for the server 102, the service provided by the server may not be a single service, for example, for the video server, the server may not be limited to only the video service, but also provide other types of multimedia services such as audio service, and for the audio server, the server 102 may also provide more types of multimedia services such as video service, and of course, the server 102 may also provide functions such as forwarding, commenting, and the like, which is not specifically limited in this embodiment of the present application. The video online playing service may refer to converting a certain movie into a video data stream, and providing the video data stream to the terminal device 101 through a video client or a web portal for online playing or offline downloading.

The server 102 is further configured to provide a bullet screen information service, which may include: multimedia resource retrieval service and barrage service. The multimedia resource retrieval service can be used in combination with the barrage service, that is, multimedia information is converted to enable the multimedia resource to correspond to barrage information in the barrage service, and a multimedia information database is provided, where the multimedia information database can be used to store information required for conversion, such as conversion rules, correspondence between multimedia information, and the like, for conversion, so as to provide accurate barrage information service for different platforms or clients, and certainly, description information of the multimedia resource itself, such as multimedia playing duration, and the like, can also be stored in the multimedia information database. The bullet screen service means that the bullet screen information server can collect bullet screen information and provide bullet screen information corresponding to the multimedia resource currently played by the client. The server 102 collects and stores the bullet screen information and the bullet screen related information, where the bullet screen related information includes a user identifier of a bullet screen sender, bullet screen sending time, a resource identifier of multimedia resources related to the bullet screen information, and the like. The user identifier may be an identifier supported by the bullet screen information server and used for uniquely identifying a bullet screen sender, the bullet screen sending time may be a time point when the user actually publishes the bullet screen content, and the multimedia resources associated with the bullet screen, that is, the resource identifier of the multimedia resources being watched by the user when the user sends the bullet screen information.

The user can access the server 102 through a client installed in the terminal apparatus 101 to use the multimedia service provided by the server 102. For example, the terminal device 101 may access the server 102 through a video client, and may also access a web portal of the server 102 through a browser client. In the process that a user watches multimedia resources through a client, the client sends a bullet screen acquisition request to the server 102 at intervals, the server 102 acquires bullet screen information of corresponding time according to the current playing progress of the client and sends the bullet screen information to the client, and the client displays the acquired bullet screen information in a playing interface. While watching the multimedia resource, the user may also input the barrage information through the client of the terminal device 101, and send a barrage sending request to the server 102, where the barrage sending request includes: after receiving the barrage sending request, the server 102 stores the information in the barrage sending request and the multimedia resource corresponding to the resource identifier in the barrage sending request in an associated manner. Meanwhile, the server 102 analyzes the barrage information sent by the user to obtain a comment object for the barrage information, further determines the interest tag of the user, and recommends the multimedia resource matched with the interest tag of the user to the user. Specifically, the method can be used for determining the interest tag of the user based on the barrage information sent by the user in real time in the process of watching the multimedia resource by the user, and recommending the content with the interest tag; or, the server may maintain an interest tag of each user, for example, collect bullet screen information sent by the user, update the interest tag of the user based on the collected bullet screen information, and recommend a multimedia resource based on the interest tag of the user when the user sends a resource acquisition request to the server, and the client may send the resource acquisition request to the server when the user logs in the client or the user manually refreshes the content of the client.

Of course, the method provided in the embodiment of the present application is not limited to be used in the application scenario shown in fig. 1, and may also be used in other possible application scenarios, and the embodiment of the present application is not limited. The functions that can be implemented by each device in the application scenario shown in fig. 1 will be described in the following method embodiments, and will not be described in detail herein.

To further illustrate the technical solutions provided by the embodiments of the present application, the following detailed description is made with reference to the accompanying drawings and the detailed description. Although the embodiments of the present application provide the method operation steps as shown in the following embodiments or figures, more or less operation steps may be included in the method based on the conventional or non-inventive labor. In steps where no necessary causal relationship exists logically, the order of execution of the steps is not limited to that provided by the embodiments of the present application.

The multimedia resource recommendation method in the embodiment of the application can be applied to the server or the terminal device shown in fig. 1, and can also be executed cooperatively by the server and the terminal device. The following describes the technical solution provided in the embodiment of the present application with reference to the application scenario shown in fig. 1.

Referring to fig. 2, an embodiment of the present application provides a multimedia resource recommendation method, including the following steps:

s201, acquiring bullet screen information sent by a user.

When a user inputs bullet screen information through a client of the terminal device and clicks a bullet screen release button, the client can display the bullet screen information on a playing interface, and meanwhile, the client can send a bullet screen sending request to the server, wherein the bullet screen sending request can comprise: bullet screen information, a user identification of the user, bullet screen sending time, a resource identification of multimedia resources related to the bullet screen information and the like. And the server receives and stores the information in the bullet screen sending request.

Taking the application to the server as an example, in step S201, the acquired bullet screen information may be the bullet screen information sent by the user through the terminal device in real time, or the historical bullet screen information sent by the user and stored in the server, and the acquired bullet screen information may be one piece of bullet screen information or multiple pieces of bullet screen information. Taking the application to the terminal device as an example, the information acquired in step S201 may be the barrage information sent by the user in real time through the terminal device, or may be one or more pieces of barrage information sent when the user watches the same multimedia resource.

In practical application, a user can send bullet screen information in a text or voice mode, and for the voice bullet screen information, the server can convert the voice bullet screen information into text bullet screen information and then perform subsequent processing based on the text bullet screen information.

S202, determining the comment object aimed at by the bullet screen information and the love degree of the user to the comment object based on the bullet screen information.

S203, based on the comment object and the love degree, determining the interest tag of the user.

In the process that a user watches multimedia resources, when the user is interested in a certain object, the barrage information aiming at the object is generally sent, and comments are made on the object, such as ' Zhang Sanzhi Fuzheng ' and ' dislike Liqu ', therefore, the object of the user comment can be extracted from the barrage information, the user's love degree on the comment object is determined based on the barrage information, and the interest tag of the user is determined based on the extracted comment object and the love degree on the comment object. For example, if the user a mentions "zhangsan" in the bullet screen information sent by the user a and likes "zhangsan", the "zhangsan" may be used as an interest tag of the user a, and then a multimedia resource related to "zhangsan" may be pushed; or, the actor lie four is mentioned in the bullet screen information sent by the user a, and the user dislikes lie four, or "actor lie four" and the user's liking of lie four as "disliking" can be used as the interest tags of the user a, and then the multimedia resources related to the actor lie four "are not pushed. Specifically, the object commented by each piece of barrage information and the love degree of the comment object by the user can be determined through semantic recognition, entity word recognition and the like, and the object commented by the barrage information can be determined by combining with multimedia resources related to the barrage information.

And S204, recommending the multimedia resources matched with the interest tags of the users to the users.

Taking the application to the terminal device as an example, the terminal device may send the interest tag determined in steps S201 to S203 to the server, the server obtains the multimedia resource matched with the interest tag and returns the multimedia resource to the terminal device, and the terminal device displays the multimedia resource matched with the interest tag returned by the server.

Taking application to a server as an example, in specific implementation, in the process of watching a certain multimedia resource by a user, the user sends barrage information, at this time, the server can determine a comment object targeted by the barrage information and the user's likeness to the comment object based on the barrage information, determine an interest tag of the user based on the comment object and the likeness to the comment object, and recommend the multimedia resource matched with the interest tag to the user, at this time, the interest tag before the user does not need to be considered, so that an interest point of the user when watching the multimedia resource can be better mined, and accurate pushing is realized.

In a specific implementation, the server may further maintain a respective interest tag set for each user, where each interest tag in the interest tag set corresponds to an object that is interested by the user, and the interest tag set may include an interest tag set by the user itself or an interest tag obtained based on analysis of behavior data of the user. When a comment object for which the barrage information sent by the user aims does not appear in the interest tag set, a new interest tag can be generated based on the comment object and added to the interest tag set of the user; when the comment object for which the barrage information sent by the user is directed exists in the interest tag set, the love degree of the user for the comment object can be determined based on the barrage information, and whether the interest tag corresponding to the comment object in the interest tag set needs to be updated or not is judged based on the love degree of the user for the comment object, for example, the love degree of the comment object changes, and then the love degree of the user for the comment object in the interest tag set can be updated. By the method, the interest label set can be continuously updated based on the bullet screen information sent by the user, the latest interest labels of the user are obtained, the interest labels of the user are enriched, the multimedia resources are recommended to the user based on the updated interest labels, and the recommendation pertinence and effectiveness are improved.

According to the multimedia resource recommendation method, the comment object aimed at by the bullet screen information and the love degree of the comment object by the user are obtained by analyzing the bullet screen information sent by the user, the interest tag of the user is determined based on the comment object aimed at by the bullet screen information and the love degree of the comment object by the user, and the multimedia resource matched with the interest tag of the user is recommended to the user. Based on the bullet screen information sent by the user, the object which the user is interested in is determined, the current interest point of the user can be more accurately positioned, the recommended multimedia resources are closer to the content which the user really is interested in, the pertinence and the effectiveness of the recommended content are improved, the use viscosity of the user is further improved, and the multimedia resource playing amount and the daily life of the user are also favorably improved.

In specific implementation, the comment object to which the bullet screen information is directed can be determined in at least one of the following ways:

the first method is to extract a comment object for which the bullet screen information is directed from the bullet screen information.

Specifically, named entity recognition may be performed on the bullet screen information, and the recognized named entity may be used as a comment object. For example, the names of actors and directors in the bullet screen information may be used as the comment objects.

Specifically, the comment object can be determined by combining the multimedia resource associated with the bullet screen information. For example, text information related to a multimedia resource associated with the barrage information may be acquired, an object that a user may comment during viewing of the multimedia resource is determined based on the acquired text information, an object set including a plurality of objects is acquired, whether an object in the object set appears in the barrage information is determined based on the object set, and an object keyword that appears is determined as a comment object to which the barrage information is directed. Taking a video as an example, text information such as a brief introduction, an actor list, a director, and a music score related to the video may be obtained, and an object set may be determined based on the text information, where the object set may include: director name, actor name, video type, etc. When the barrage information sent by the user contains the director name, the director name can be identified based on the object set, and then the director name is used as a comment object.

In practical application, taking a video as an example, the object set corresponding to the video may further include a correspondence between the roles and the actors, and at this time, the comment object may be identified more accurately according to the correspondence between the roles and the actors. For example, the bullet screen information sent by the user when viewing the video is "actor skill bar", the server may first recognize the named entity "actor" in the bullet screen information, and then determine that the actor corresponding to the "actor" is lie four based on the correspondence between the role of the video and the actor, and then take the lie four as the comment object to which the bullet screen information is directed. Thus, when some pronouns such as "female owner", "male owner", "first male number" and the like appear in the bullet screen information, the comment object targeted by the bullet screen information can still be accurately identified.

And in the second mode, the target multimedia resource associated with the bullet screen information is determined as the comment object aimed at by the bullet screen information.

In specific implementation, when storing the bullet screen information, the server associates and stores the resource identifier of the target multimedia resource corresponding to the bullet screen information, that is, the bullet screen information is sent when the user views the target multimedia resource. Therefore, the target multimedia resource associated with the bullet screen information can be determined through the resource identifier corresponding to the bullet screen information, and the target multimedia resource is determined as the comment object.

For example, when the barrage information is barrage information transmitted when the user watches video a, video a may be determined as a comment object to which the barrage information is directed.

In specific implementation, the user's preference for the comment object targeted by each bullet screen information can be obtained from each bullet screen information based on semantic recognition or keyword matching and other modes. Wherein the likeness can be described by at least two emotion ratings, for example, the likeness can be divided into two emotion ratings like and dislike, or the likeness can be divided into five emotion ratings like very much, like generally, neutral, dislike generally and dislike very much.

Taking keyword matching as an example, an emotion degree keyword list corresponding to each emotion level describing a like degree can be preset, for example, the like corresponding emotion degree keywords include: "really nice looking", "too excellent", "skill cracking", "too wonderful", etc., dislike the corresponding emotional keywords including: "too rotten", "unsightly", "nausea", etc.

To this end, the user's likeness to the comment object may be determined as follows: extracting emotion degree keywords from the bullet screen information; and determining the love degree of the user to the comment object based on the extracted emotional degree key words. Then, based on the user's liking of the comment object, an interest tag for identifying the user's liking of the comment object is obtained.

Taking the example that the love degree comprises two emotion levels of like and dislike, if the barrage information sent by the user comprises a keyword of the emotion degree corresponding to the like, determining that the love degree of the comment object aimed at by the barrage information is like by the user; and if the bullet screen information sent by the user comprises the emotional degree key words which do not like the corresponding emotional degree key words, determining that the user does not like the comment object which is aimed at by the bullet screen information. Then, based on the user's liking of the comment object, an interest tag for identifying the user's liking of the comment object is obtained.

Taking semantic recognition as an example, an emotion recognition model can be trained, and based on the emotion recognition model, the love of the user on the comment object is determined. The input of the training emotion recognition model is barrage information, and the output is emotion level representing the love. Taking the example that the likeness can be divided into two emotion grades like and dislike, the emotion recognition model can be a two-class neural network, and the two-class deep neural network includes, but is not limited to, CNN, RNN, LSTM, DNN, DBNs, etc. And acquiring a large amount of texts marked with emotion levels as training samples, and training the two-classification deep neural network based on the training samples to obtain an emotion recognition model capable of recognizing the emotion levels of the input texts. And inputting the barrage information into a trained emotion recognition model to obtain the emotion level corresponding to the barrage information, and taking the emotion level as the love of the user on the comment object aimed at by the barrage information. Through the emotion recognition model, some recessive expression like or dislike modes can be recognized, and the recognition accuracy and recognition efficiency of the likeness are improved.

By the mode, the love degree of the user to the comment object contained in the bullet screen information can be determined based on any bullet screen information sent by the user, and then the interest tag used for identifying the love degree of the user to the comment object is obtained.

When the comment object is extracted from the bullet screen information, that is, the comment object is extracted in the first way, at this time, the user's liking for the comment object can be determined in any one of the following ways: extracting emotion keywords from the barrage information, and determining the love of the user on the comment object based on the extracted emotion keywords; or inputting the barrage information into the emotion degree recognition model to obtain the love degree of the user on the comment object in the barrage information. Then, based on the user's liking of the comment object, an interest tag for identifying the user's liking of the comment object is obtained.

In particular, only the forward interest tag may be set, that is, only the interest tag is set for the object that the user likes. In this case, when recommendation is performed based on the interest tag, a multimedia resource including an object to which the interest tag of the user is directed is selected and recommended to the user. For example, if the interest tags of the user include "actor zhang san" and "suspense," a suspense-like movie in which actor zhang san is played may be recommended to the user.

When only the forward interest tag is set, if the user's likeness to the comment object meets the preset condition, the comment object or the category to which the comment object belongs is determined as the interest tag of the user. The preset condition may be that a preset emotion level is reached, for example, the preset emotion level is "like", or the preset emotion levels are "like very much" and "like generally". The category to which the comment object belongs may be determined according to an actual application scenario, for example, for a video, the category may be a video type, such as suspicion, love, story, and the like.

For example, the preset emotion level is "like", and when the user's likeness to the comment object "three actors" is "like", the "three actors" is used as the interest tag of the user. When the user's likeness to the comment object "director lie four" is "dislike", the "director lie four" is not taken as the interest tag of the user.

For another example, the preset emotion level is "favorite", and when the user's likeness to the comment object "movie B" is "favorite", the movie genre "suspense" to which "movie B" belongs is used as the interest tag of the user.

Of course, a positive interest tag and a negative interest tag may also be set at the same time, where the positive interest tag is an interest tag set for an object that the user likes, and the negative interest tag is an interest tag set for an object that the user dislikes. Under the above circumstances, when recommendation is performed based on the interest tags, multimedia resources including objects that are not liked by the user in the first candidate multimedia resource set may be filtered based on the negative interest tags to obtain a second candidate multimedia resource set, and then multimedia resources including objects that are liked by the user may be selected from the second candidate multimedia resource set based on the positive interest tags and recommended to the user. For example, if the positive interest tags of the user include "movie in casting three and" movie in suspense "and the negative interest tags include" movie in suspense "and" movie in director's lie four ", the movie in suspense with casting three and showing three may be recommended to the user, but the recommended movie does not include the movie taken by" movie in director's lie four ".

When a positive interest tag and a negative interest tag are set at the same time, if the user's love degree on the comment object meets a positive preset condition, determining the comment object or the category to which the comment object belongs as the positive interest tag of the user; if the user's love to the comment object meets a negative preset condition, determining the comment object or the category to which the comment object belongs as a negative interest tag of the user; for the case that neither the positive preset condition nor the negative preset condition is satisfied, the interest tag may be selected not to be generated.

For example, the forward preset condition is: the emotion level is 'very favorite', and the negative preset condition is as follows: the emotional level is "very dislike". When the user's liking degree of the comment object "movie B" is "very liking", the movie type "suspense movie" to which "movie B" belongs is taken as the forward interest tag of the user; when the user's liking degree of the comment object ' director lie four ' is ' very dislike ', the ' director lie four ' is used as a negative interest tag of the user; when the user's likeness to the comment object "movie C" is "general like", the interest tag is not generated.

When the comment object is extracted from the bullet screen information, that is, the comment object is extracted in the second manner, referring to fig. 3, the interest tag for identifying the user's liking of the comment object can be obtained in the following manner:

s301, aiming at each bullet screen information which is sent by a user and is related to the target multimedia resource, determining the emotion level corresponding to each bullet screen information based on the emotion degree key words contained in each bullet screen information.

In specific implementation, the emotion degree keyword can be extracted from the bullet screen information based on modes such as keyword matching and the like, and the emotion level corresponding to the bullet screen information is determined based on the extracted emotion degree keyword.

In specific implementation, the emotion level corresponding to each piece of bullet screen information can be determined based on the emotion degree identification model.

S302, determining the user' S preference degree for the target multimedia resource based on the emotion level corresponding to each bullet screen information.

The bullet screen information in step S302 is the bullet screen information associated with the target multimedia resource sent by the user in step S301.

In specific implementation, the statistical characteristic value of the emotion level of each bullet screen information can be obtained, and the user preference degree of the target multimedia resource is determined based on the statistical characteristic value. The statistical characteristic value may be an average value, a mode, a median, or the like.

For example, when two emotion levels of like and dislike are included, the number n of bullet screen information whose emotion level is like may be counted₁And the number n of the bullet screen information with the emotional level being disliked₂If n is₁>n₂Then determining the user's liking for the target multimedia resource, if n₁<n₂Then determining the user's liking of the target multimedia resource as liking, for n₁＝n₂Can determine that the user's preference for the target multimedia resource is neutral,i.e. the attitude of the user to the target multimedia asset is neither liked nor disliked.

For another example, when five emotion levels are included, the emotion level including the largest number of bullet screen information may be determined as the user's preference for the target multimedia resource by the number of bullet screen information belonging to each emotion level. Or, each emotion level corresponds to a score, for example, "very like" corresponds to 5 scores, "general like" corresponds to 4 scores, "neutral" corresponds to 3 scores, "general dislike" corresponds to 2 scores, "very dislike" corresponds to 1 score, an average value of the scores of the emotion levels of all the bullet screen information is calculated, the emotion level corresponding to the score closest to the average value is determined as the user's preference for the target multimedia resource, for example, the average value is 4.1 scores, and the user's preference for the target multimedia resource is "general like".

Further, in the barrage information sent by the user, some comment objects of the barrage information are directly target multimedia resources, such as barrage information like "this drama is really good and" this movie is too good ", at this time, the proportion of the emotion level of this type of barrage information in the statistical characteristic value can be increased, for example, when calculating the average value, the weight of this type of barrage information is greater than 1, and the weight of other barrage information is 1.

S303, obtaining an interest tag used for identifying the user 'S preference for the comment object based on the user' S preference for the target multimedia resource.

In a specific embodiment, the manner for obtaining the interest tag for identifying the user's preference for the comment object based on the user's preference for the comment object may be referred to, and is not repeated.

And counting the overall emotion degree of the user when the user watches the target multimedia resource based on the emotion level of the barrage information sent by the user when the user watches the target multimedia resource, further acquiring the love degree of the user on the whole target multimedia resource, and determining the interest tag of the user.

Further, the method of the embodiment of the present application further includes the following steps: and determining the bullet screen participation degree when the user watches the target multimedia resource based on the quantity of the bullet screen information which is sent by the user and is associated with the target multimedia resource.

The bullet screen participation degree is an index used for measuring the positivity of sending the bullet screen when the user watches the target multimedia resource, and the higher the bullet screen participation degree is, the more interesting the user is in the target multimedia resource.

In a possible implementation manner, the number of the barrage information related to the target multimedia resource sent by the user is positively correlated to the barrage participation degree, that is, the more the barrage information sent by the user when viewing the target multimedia resource, the higher the barrage participation degree of the user is.

In specific implementation, the number M of barrage information associated with the target multimedia resource sent by the user can be based on₁And the average bullet screen sending number M of the users₂And determining the bullet screen participation degree when the user watches the target multimedia resource. Wherein, the average bullet screen sending number M₂The number of the barrage information sent when the user watches other multimedia resources is determined, for example, the number of the barrage information sent when the user watches N multimedia resources recently can be obtained, and the average value of the N numbers is taken as the average barrage sending number M₂And the value of N can be set according to actual requirements. In particular, according to the number M₁And average bullet screen sending number M₂The bullet screen participation degree is determined according to the ratio, and the bullet screen participation degree can also be determined according to the quantity M₁And average bullet screen sending number M₂Determining the bullet screen participation degree. Sending number M of average bullet screens of users₂As a reference, the bullet screen participation of the user is determined, and the habit of sending bullet screens by the user is fully considered, for example, some users send many bullet screens regardless of whether they like the watched multimedia resources, and some users send more bullet screens only when they watch the favorite multimedia resources.

In another possible implementation manner, the bullet screen sending frequency when the user watches the target multimedia resource can be determined based on the quantity of bullet screen information which is sent by the user and is associated with the target multimedia resource; and determining the bullet screen participation degree when the user watches the target multimedia resource based on the bullet screen sending frequency. Wherein, the bullet screen sending frequency is positively correlated with the bullet screen participation degree. For example, the number of barrages transmitted in a unit time when the user views the target multimedia resource, for example, the average number of barrages transmitted in one minute, may be counted, and the number of barrages transmitted in a unit time may be used as the barrage transmission frequency when the user views the target multimedia resource.

In specific implementation, the barrage participation degree of the user watching the target multimedia resource can be determined based on the barrage sending frequency and the historical barrage sending frequency of the user. The historical bullet screen sending frequency of the user is determined according to the bullet screen sending frequency when the user watches other multimedia resources, for example, the bullet screen sending frequency when the user watches N multimedia resources recently can be obtained, the average value of the N bullet screen sending frequencies is calculated and used as the historical bullet screen sending frequency, and the value of N can be set according to actual requirements. Specifically, the bullet screen participation degree can be determined according to the ratio of the bullet screen sending frequency to the historical bullet screen sending frequency, and the bullet screen participation degree can also be determined according to the difference value of the bullet screen sending frequency and the historical bullet screen sending frequency. The method comprises the steps of determining the bullet screen participation degree of a user by taking the historical bullet screen sending frequency of the user as a reference, and fully considering the habit of sending bullet screens of the user, for example, some users can send a lot of bullet screens regardless of whether the users like watched multimedia resources, and some users send more bullet screens only when the users watch the favorite multimedia resources.

On this basis, step S302 specifically includes: and determining the user's preference degree for the target multimedia resource based on the emotion level and the bullet screen participation degree corresponding to each bullet screen information.

During specific implementation, the overall emotion degree of the user on the target multimedia resource can be determined based on the emotion level corresponding to each bullet screen information. For example, a statistical feature value of the emotion level of each bullet screen information may be obtained, and the emotion degree of the target multimedia resource is determined based on the statistical feature value, where the statistical feature value may be an average value, a mode, a median, and the like, and the specific embodiment may refer to step S302. Then, based on the overall emotion and barrage participation of the user on the target multimedia resource, the love of the user on the target multimedia resource is determined, for example, the overall emotion and barrage participation can be quantized into corresponding scores respectively, the score corresponding to the overall emotion and the score corresponding to the barrage participation are weighted and summed to obtain a weighted score, and the love corresponding to the weighted score is determined.

The method and the device determine the liking degree of the user to the target multimedia resource by combining the barrage participation degree of the barrage sending and the whole emotion degree of the target multimedia resource when the user watches the target multimedia resource, and improve the accuracy rate of calculating the liking degree of the user to the target multimedia resource.

Based on the same inventive concept as the multimedia resource recommendation method, an embodiment of the present application further provides another multimedia resource recommendation method, which can be applied to a terminal device, and with reference to fig. 4, the multimedia resource recommendation method specifically includes the following steps:

401. and receiving bullet screen information sent by a user.

When a user inputs bullet screen information through a playing interface of the terminal equipment and clicks a bullet screen release button, the terminal equipment receives the bullet screen information input by the user and displays the bullet screen information on the playing interface, meanwhile, the terminal equipment sends a bullet screen sending request to the server, the bullet screen sending request comprises the bullet screen information, and in addition, the bullet screen sending request also can comprise a user identifier of the user, bullet screen sending time, a resource identifier of multimedia resources related to the bullet screen information and the like. And the server receives and stores the information in the bullet screen sending request.

S402, displaying a multimedia resource recommendation list determined based on the barrage information, wherein multimedia resources included in the multimedia resource recommendation list are matched according to interest tags of the user, and the interest tags of the user are determined based on the comment object targeted by the barrage information and the love degree of the user on the comment object.

The multimedia resource recommendation list comprises at least one multimedia resource.

In specific implementation, a user sends a barrage information sending request to a server through terminal equipment, the server obtains barrage information in the barrage information sending request, determines a comment object for the barrage information and the love degree of the user on the comment object, determines an interest tag of the user based on the comment object and the love degree of the user on the comment object, determines to recommend multimedia resources matched with the interest tag of the user to the user based on the interest tag, and returns the multimedia resources to the terminal equipment in a multimedia resource recommendation list mode. And the terminal equipment displays the multimedia resource recommendation list returned by the server on a playing interface. The server determines the interest tag of the user based on the bullet screen information sent by the user, and refers to the multimedia resource recommendation methods shown in fig. 2 and fig. 3 for a specific implementation mode of multimedia resource recommendation based on the interest tag, which is not described again.

In specific implementation, after acquiring barrage information input by a user, the terminal device can determine a comment object aimed at by the barrage information and the user's likeness to the comment object, determine the interest tag of the user based on the comment object and the user's likeness to the comment object, send the determined interest tag to the server, acquire multimedia resources matched with the interest tag by the server, and return the multimedia resources to the terminal device, and the terminal device displays the multimedia resources matched with the interest tag returned by the server.

In specific implementation, recommendation can be performed based on one piece of bullet screen information sent by the user, and recommendation can also be performed based on a plurality of pieces of bullet screen information sent by the user when the user watches the target multimedia resource.

When recommending is performed based on one piece of bullet screen information, the server or the terminal device may obtain the bullet screen information in the bullet screen sending request, determine the comment object to which the bullet screen information corresponds, and recommend the multimedia resource matched with the comment object to the user, for example, if the comment object to which the bullet screen information sent by the user corresponds is "three cast by actor", the multimedia resource related to "three cast by actor" may be recommended to the user. Or the server may obtain the bullet screen information in the bullet screen sending request, determine the interest tag of the user based on the comment object targeted by the bullet screen information and the user's preference for the comment object, and recommend the multimedia resource matched with the interest tag of the user to the user, for example, if three actors are mentioned in the bullet screen information sent by the user, the three actors can be used as the interest tag of the user, and then recommend the multimedia resource related to the interest tag to the user.

Referring to fig. 5, in the process of watching a video, the user sends bullet screen information 501: "Zhang Sanchang Tai excellent! After receiving the bullet screen information, the server determines that the comment object to which the bullet screen information is directed is "zhangsan", which indicates that the user is interested in "zhangsan", and can immediately recommend the multimedia resource related to "zhangsan" to the user, and push the multimedia resource to the terminal device of the user in the form of a multimedia resource recommendation list, and the terminal device can display the content in the multimedia resource recommendation list on the play interface 50. Therefore, the current interest point of the user can be determined based on the comment object aimed at by the bullet screen information sent by the user in real time, and then the user can be recommended in a targeted manner, so that the real-time performance and the accuracy of recommendation are improved.

In specific implementation, the multimedia resource recommendation list may be displayed in a play interface in a form of a popup window or the like, the specific display form may refer to fig. 5, the terminal device may display the popup window 502 for displaying the multimedia resource recommendation list in the play interface 50, and display a collection control 503 and an exit control 504 for the multimedia resource recommendation list in the popup window 502. If the user is interested in the recommended multimedia resource, the favorite control 503 can be clicked, the terminal device responds to the triggering operation aiming at the favorite control 503, the multimedia resource in the multimedia resource recommendation list is added to the resource favorite of the user, the popup window 502 is closed immediately, and then the user can acquire the multimedia resource in the multimedia resource recommendation list from the resource favorite. If the user is not interested in the recommended multimedia resource, the exit control 504 can be clicked, the terminal device responds to the triggering operation aiming at the exit control 504, and the pop-up window 502 is closed. Certainly, if the preset duration is exceeded, the user still does not click the collection control 503 or exit the collection control 504, and the terminal device may automatically close the pop-up window 502, so as to avoid the pop-up window affecting the user to view the multimedia resource.

Furthermore, the multimedia resources in the multimedia resource recommendation list returned by the server are the multimedia resources matched with the favorite comment objects of the user sending the barrage information. The comment object favored by the user is determined based on the user's liking degree of the comment object for which the barrage information is directed, for example, when the emotion level of the user's liking degree of the comment object is 'liking', it is determined that the user likes the comment object, and at this time, the server can recommend the multimedia resource matched with the comment object to the user.

Specifically, the user's preference for the comment object may be determined based on the bullet screen information sent by the user in real time. The user's preference for the comment object may also be determined based on the user's historical interest tags. The method may further include determining a preference of the user for the comment object by integrating the barrage information sent by the user in real time and the historical interest tag of the user, for example, performing weighted summation on a first preference determined based on the barrage information sent by the user in real time and a second preference of the comment object by the user in the historical interest tag, and taking a weighted result as the preference of the user for the comment object, where the weights of the first preference and the second preference may be determined according to actual requirements, for example, the weight of the first preference is greater than the weight of the second preference. Meanwhile, the historical interest labels of the user can be updated based on the weighted results of the first likeness and the second likeness. The specific implementation of determining the user's preference for the comment object based on the bullet screen information may refer to the multimedia resource recommendation method shown in fig. 2 and 3, and is not described again.

For example, the user sends the barrage information "Zhang san Zhang Tai excellent! And determining that the user likes a comment object 'zhang san' based on the bullet screen information, and recommending the media resource related to the 'zhang san' to the user by the server. When a user watches a video, the user sends barrage information that the Li Si Tu is poor in reality! If it is determined that the user does not like the comment object "lie four" based on the bullet screen information, the server does not recommend the media resource related to "lie four" to the user.

When recommendation is performed based on a plurality of pieces of bullet screen information, after receiving N pieces of bullet screen information sent when a user watches a target multimedia resource, a server or terminal equipment respectively determines a comment object for each piece of bullet screen information in the N pieces of bullet screen information; then, based on a plurality of pieces of bullet screen information aiming at the same comment object, the love degree of the user on the comment object is determined, and whether the multimedia resource matched with the comment object is recommended to the user or not is determined based on the love degree of the user on the comment object.

Alternatively, a sampling period may be set, which may be determined based on the play duration of the target multimedia asset, for example, the sampling period may be 1/N of the play duration, and the target multimedia asset refers to a multimedia asset currently being viewed by the user. The method comprises the steps that a server or terminal equipment obtains bullet screen information sent by a user in a sampling period, and a comment object corresponding to each piece of bullet screen information in the bullet screen information is respectively determined; then, based on a plurality of pieces of bullet screen information aiming at the same comment object, the love degree of the user on the comment object is determined, and whether the multimedia resource matched with the comment object is recommended to the user or not is determined based on the love degree of the user on the comment object.

In specific implementation, the overall emotion degree of the user watching the target multimedia resource can be obtained through statistics based on the emotion grades of a plurality of barrage information sent by the user watching the target multimedia resource, so that the preference degree of the user on the whole target multimedia resource is obtained, and if the emotion grade of the preference degree of the user on the whole target multimedia resource is determined to be preferred, the multimedia resource matched with the target multimedia resource can be recommended to the user. The method can determine the liking degree of the user to the whole target multimedia resource based on a plurality of barrage information sent by the user when the user watches the target multimedia resource after the user watches the target multimedia resource or the target multimedia resource is about to be played, so as to improve the accuracy of liking degree calculation, and send the multimedia resource matched with the target multimedia resource to the terminal equipment when the target multimedia resource is played or is about to be played. The specific implementation of obtaining the user's preference for the entire target multimedia resource based on the emotion levels of the multiple pieces of bullet screen information may refer to the method in fig. 3, and is not described again.

Certainly, for the scene based on the real-time recommendation of the bullet screen information, each piece of bullet screen information sent by the user can not be pushed, so that the watching experience of the user can be prevented from being influenced by too many pushes. For example, if the comment object cannot be extracted from the bullet screen information, the multimedia resource may not be pushed. For another example, in the process of watching the same multimedia resource, the server already recommends a multimedia resource related to a certain comment object to the user, and then does not recommend the comment object any more, even if the user sends the barrage information for the comment object again. Or, within a time period, after a multimedia resource related to a certain comment object has been recommended to the user, the comment object is not recommended any more, even if the user sends the barrage information for the comment object again, the time period may be set according to actual needs, for example, the time period may be 1 hour, 1 day, a week, or the like.

For this reason, the server may record a history recommended object list of the user, the history recommended object list including a comment object recommended to the user, and if a comment object specified based on the bullet screen information exists in the history recommended object list of the user, the server may not recommend the comment object. The history recommended object list further includes a recommended time of each comment object, and the server may determine whether the comment object needs to be deleted from the history recommended object list based on the recommended time of the comment object, specifically, may set an update period, where the update period may be a specific time, such as 1 hour, 1 day, or one week. Of course, the user's historical recommendation object list may also be emptied immediately after the user finishes watching the current multimedia resource.

The multimedia resource recommendation method taking fig. 4 as an example can determine the comment object to which the barrage information is directed based on the barrage information sent in real time in the process of watching the multimedia resources by the user, thereby determining the current interest point of the user, and further performing targeted recommendation on the multimedia resources to the user in the process of watching the multimedia resources by the user, so as to improve the real-time performance and accuracy of the recommendation.

Based on the same inventive concept as the multimedia resource recommendation method, an embodiment of the present application further provides another multimedia resource recommendation method, which can be applied to a terminal device, and with reference to fig. 6, the multimedia resource recommendation method specifically includes the following steps:

s601, sending a resource obtaining request to a server, wherein the resource obtaining request comprises a user identifier, so that the server returns a corresponding multimedia resource recommendation list, the multimedia resource recommendation list comprises multimedia resources matched with interest tags corresponding to the user identifier, and the interest tags are determined based on comment objects corresponding to bullet screen information sent by a user corresponding to the user identifier.

The determination method of the interest tag corresponding to each user identifier may refer to a multimedia resource recommendation method on the server side, and is not described again.

And S602, receiving and displaying the multimedia resource recommendation list returned by the server.

In specific implementation, the resource acquisition request may be initiated manually by the user, or may be initiated automatically by the client at a suitable time.

For example, the terminal device may send a resource acquisition request to a server corresponding to the client when the user opens the client or selects to acquire recommended content; the server searches the interest tag corresponding to the user identifier based on the user identifier in the resource acquisition request, screens out the multimedia resource matched with the interest tag corresponding to the user identifier, generates a multimedia resource recommendation list and returns the multimedia resource recommendation list to the terminal equipment; and the terminal equipment displays the multimedia resources in the multimedia resource recommendation list in a display interface of the client for a user to view.

For example, the terminal device may also send a resource acquisition request to a server corresponding to the client when the user refreshes the multimedia resource recommendation list of the client, so as to reacquire other recommended multimedia resources; the server searches the interest tag corresponding to the user identifier based on the user identifier in the resource acquisition request, screens out the multimedia resource matched with the interest tag corresponding to the user identifier from the media resource which is not recommended to the user, generates a multimedia resource recommendation list and returns the multimedia resource recommendation list to the terminal equipment; and the terminal equipment displays the multimedia resources in the multimedia resource recommendation list in a display interface of the client for a user to view.

The multimedia resource recommendation method taking fig. 6 as an example can analyze bullet screen information sent by a user to obtain a comment object targeted by the bullet screen information, determine an interest tag of the user based on the comment object targeted by the bullet screen information, improve the positioning accuracy of the interest tag, and recommend the user based on the interest tag of the user when the user needs to be recommended, so that the recommended multimedia resource is closer to the content really interested by the user, the pertinence and the validity of the recommended content are improved, the use viscosity of the user is further improved, and the multimedia resource playback amount and the daily life of the user are also favorably improved.

As shown in fig. 7, based on the same inventive concept as the multimedia resource recommendation method, the embodiment of the present application further provides a multimedia resource recommendation device 70, which includes an obtaining module 701, an information extracting module 702, an interest determining module 703 and a recommending module 704.

An obtaining module 701, configured to obtain bullet screen information sent by a user;

the information extraction module 702 is configured to determine, based on the bullet screen information, a comment object targeted by the bullet screen information and a user's preference for the comment object;

an interest determination module 703, configured to determine an interest tag of the user based on the comment object and the likeness;

and a recommending module 704 for recommending the multimedia resource matched with the interest tag of the user to the user.

Optionally, the information extraction module 702 is specifically configured to determine the comment object to which the bullet screen information is directed by at least one of the following manners:

extracting a comment object aimed at by the bullet screen information from the bullet screen information; or,

Optionally, when the comment object is extracted from the bullet screen information, the information extraction module 702 is specifically configured to determine the user's preference for the comment object by:

extracting emotion degree keywords from the bullet screen information;

Optionally, when the comment object is determined according to the target multimedia resource associated with the barrage information, the information extraction module 702 is specifically configured to determine the user's likeness to the comment object by:

determining an emotion level corresponding to each bullet screen information based on an emotion degree keyword contained in each bullet screen information aiming at each bullet screen information which is sent by a user and is associated with a target multimedia resource;

Optionally, the information extracting module 702 is further configured to determine a bullet screen participation degree when the user watches the target multimedia resource based on the number of bullet screen information sent by the user and associated with the target multimedia resource;

the information extraction module 702 is specifically configured to determine a user's preference for the target multimedia resource based on the emotion level and the bullet screen participation degree corresponding to each piece of bullet screen information.

Optionally, the information extracting module 702 is specifically configured to:

and determining the bullet screen participation degree when the user watches the target multimedia resource based on the bullet screen sending frequency.

The multimedia resource recommendation device and the multimedia resource recommendation method provided by the embodiment of the application adopt the same inventive concept, can obtain the same beneficial effects, and are not repeated herein.

As shown in fig. 8, based on the same inventive concept as the multimedia resource recommendation method, the embodiment of the present application further provides a multimedia resource recommendation device 80, which includes a first receiving module 801 and a display module 802.

A first receiving module 801, configured to receive bullet screen information sent by a user;

the display module 802 is configured to display a multimedia resource recommendation list determined based on the barrage information, where multimedia resources included in the multimedia resource recommendation list are matched according to interest tags of the user, and the interest tags of the user are determined based on the comment object targeted by the barrage information and the user's preference for the comment object.

As shown in fig. 9, based on the same inventive concept as the multimedia resource recommendation method, the embodiment of the present application further provides a multimedia resource recommendation device 90, which includes a sending module 901 and a second receiving module 902.

A sending module 901, configured to send a resource obtaining request to a server, where the resource obtaining request includes a user identifier, so that the server returns a corresponding multimedia resource recommendation list, where the multimedia resource recommendation list includes multimedia resources matched with interest tags corresponding to the user identifier, and the interest tags are determined based on comment objects to which bullet screen information sent by a user corresponding to the user identifier is directed and likeliness of the comment objects by the user;

and a second receiving module 902, configured to receive and display the multimedia resource recommendation list returned by the server.

Based on the same inventive concept as the multimedia resource recommendation method, an embodiment of the present application further provides an electronic device, which may specifically be a server or a terminal device shown in fig. 1. As shown in fig. 10, the electronic device 100 may include a processor 1001 and a memory 1002.

The Processor 1001 may be a general-purpose Processor, such as a Central Processing Unit (CPU), a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, a discrete Gate or transistor logic device, or a discrete hardware component, and may implement or execute the methods, steps, and logic blocks disclosed in the embodiments of the present Application. A general purpose processor may be a microprocessor or any conventional processor or the like. The steps of a method disclosed in connection with the embodiments of the present application may be directly implemented by a hardware processor, or may be implemented by a combination of hardware and software modules in a processor.

Memory 1002, which is a non-volatile computer-readable storage medium, may be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The Memory may include at least one type of storage medium, and may include, for example, a flash Memory, a hard disk, a multimedia card, a card-type Memory, a Random Access Memory (RAM), a Static Random Access Memory (SRAM), a Programmable Read Only Memory (PROM), a Read Only Memory (ROM), a charged Erasable Programmable Read Only Memory (EEPROM), a magnetic Memory, a magnetic disk, an optical disk, and so on. The memory is any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer, but is not limited to such. The memory 1002 in the embodiments of the present application may also be circuitry or any other device capable of performing a storage function for storing program instructions and/or data.

An embodiment of the present application provides a computer-readable storage medium for storing computer program instructions for the electronic device, which includes a program for executing the multimedia resource recommendation method.

The computer storage media may be any available media or data storage device that can be accessed by a computer, including but not limited to magnetic memory (e.g., floppy disks, hard disks, magnetic tape, magneto-optical disks (MOs), etc.), optical memory (e.g., CDs, DVDs, BDs, HVDs, etc.), and semiconductor memory (e.g., ROMs, EPROMs, EEPROMs, non-volatile memory (NAND FLASH), Solid State Disks (SSDs)), etc.

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

The above embodiments are only used to describe the technical solutions of the present application in detail, but the above embodiments are only used to help understanding the method of the embodiments of the present application, and should not be construed as limiting the embodiments of the present application. Modifications and substitutions that may be readily apparent to those skilled in the art are intended to be included within the scope of the embodiments of the present application.

Claims

1. A multimedia resource recommendation method is characterized by comprising the following steps:

acquiring bullet screen information sent by a user;

2. The method according to claim 1, wherein the determining of the comment object for which the barrage information is directed specifically includes at least one of:

3. The method according to claim 2, characterized in that when the comment object is extracted from the bullet screen information, the user's likeability to the comment object is determined by:

extracting emotion keywords from the bullet screen information;

4. The method of claim 2, wherein when the comment object is determined according to the target multimedia resource associated with the barrage information, the user's likeness to the comment object is determined by:

5. The method of claim 4, further comprising:

determining the bullet screen participation degree of the user when the user watches the target multimedia resource based on the quantity of the bullet screen information which is sent by the user and is related to the target multimedia resource;

the determining the user's preference for the target multimedia resource based on the emotion level corresponding to each bullet screen information specifically includes:

and determining the user's preference for the target multimedia resource based on the emotion level corresponding to each bullet screen information and the bullet screen participation.

6. The method according to claim 5, wherein the determining the bullet screen engagement degree when the user views the target multimedia resource based on the number of bullet screen information associated with the target multimedia resource sent by the user specifically comprises:

7. A multimedia resource recommendation method is characterized by comprising

Receiving bullet screen information sent by a user;

8. A multimedia resource recommendation method is characterized by comprising the following steps:

9. A multimedia resource recommendation apparatus, comprising:

10. A multimedia resource recommendation apparatus, comprising:

11. A multimedia resource recommendation apparatus, comprising:

12. An electronic device comprising a memory, a processor and a computer program stored on the memory and executable on the processor, characterized in that the steps of the method of any of claims 1 to 8 are implemented when the computer program is executed by the processor.

13. A computer-readable storage medium having computer program instructions stored thereon, which, when executed by a processor, implement the steps of the method of any one of claims 1 to 8.