CN111818370B

CN111818370B - Information recommendation method and device, electronic equipment and computer-readable storage medium

Info

Publication number: CN111818370B
Application number: CN202010620563.6A
Authority: CN
Inventors: 孔凡阳
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2020-06-30
Filing date: 2020-06-30
Publication date: 2022-03-11
Anticipated expiration: 2040-06-30
Also published as: CN111818370A

Abstract

The invention provides an information recommendation method, an information recommendation device, electronic equipment and a computer-readable storage medium; the method comprises the following steps: presenting interactive information aiming at the video in the video playing process; performing semantic analysis processing on the interaction information to obtain the interest degree of the interaction information representation aiming at the object in the video; determining recommendation information according to the interest degree of the interaction information representation aiming at the object in the video; and presenting the recommendation information. By the method and the device, accurate recommendation can be realized on the basis that the user watches videos.

Description

Information recommendation method and device, electronic equipment and computer-readable storage medium

Technical Field

The present invention relates to internet technologies, and in particular, to an information recommendation method and apparatus, an electronic device, and a computer-readable storage medium.

Background

Artificial intelligence is a theory, method and technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use the knowledge to obtain the best results. Artificial intelligence is now rapidly developing and widely used in various industries.

Taking an application scenario of information recommendation as an example, in the related art, corresponding recommendation information is generally configured in advance in a video, so that the preconfigured recommendation information is recommended to a user in a process of watching the video by the user. Since the pre-configured recommendation information is fixed and not in place, the user is not necessarily interested in it, which may result in a low probability that the pre-configured recommendation information is clicked by the user, and thus lack of exposure drainage.

Disclosure of Invention

The embodiment of the invention provides an information recommendation method and device, electronic equipment and a computer readable storage medium, which can realize accurate recommendation on the basis of watching videos by users.

The technical scheme of the embodiment of the invention is realized as follows:

the embodiment of the invention provides an information recommendation method, which comprises the following steps:

presenting interactive information aiming at the video in the video playing process;

performing semantic analysis processing on the interaction information to obtain the interest degree of the interaction information representation aiming at the object in the video;

determining recommendation information according to the interest degree of the interaction information representation aiming at the object in the video;

and presenting the recommendation information.

An embodiment of the present invention provides an information recommendation apparatus, including:

the interactive presentation module is used for presenting interactive information aiming at the video in the video playing process;

the semantic analysis module is used for performing semantic analysis processing on the interaction information to obtain the interest degree of the interaction information representation aiming at the object in the video;

the recommendation module is used for determining recommendation information according to the interest degree of the interaction information representation aiming at the object in the video; and presenting the recommendation information.

In the above scheme, the interactive presentation module is further configured to present a video interactive interface; and responding to the interactive operation aiming at the video received in the video interactive interface, and presenting the interactive information corresponding to the interactive operation.

In the above scheme, the recommendation module is further configured to determine, as the recommendation information, candidate recommendation information whose degree of correlation with an object in the video satisfies a correlation condition; wherein the degree of correlation characterizes the correlation between the object recommended by the recommendation information and the object in the video; the degree of correlation has a positive correlation with the degree of interest.

In the above scheme, the recommendation module is further configured to select at least one piece of recommendation information in a descending order when the interaction information is interested in the degree of interest of the object in the video; the descending order is obtained by sorting the plurality of candidate recommendation information according to the correlation degree from high to low; when the interest degree of the interaction information aiming at the object in the video is uninteresting, selecting at least one piece of previous recommendation information in ascending order; the ascending order is obtained by sorting the plurality of candidate recommendation information according to the correlation degree from low to high.

In the above scheme, the recommendation module is further configured to determine candidate recommendation information in the recommendation information types corresponding to the interest degree as the recommendation information; the different interest degrees correspond to the different recommendation information types, and positive correlation exists between the interest degrees and the correlation degrees, and the correlation degrees represent the correlation between objects recommended by the recommendation information in the recommendation information types and the objects in the video.

In the above scheme, the recommendation module is further configured to select at least one piece of recommendation information corresponding to the object and/or select at least one piece of recommendation information corresponding to a related object when the interaction information is interested in the degree of interest of the object in the video; wherein the related object and the object belong to the same recommendation information type; when the interest degree of the interaction information for the object in the video is uninteresting, selecting at least one piece of recommendation information corresponding to the irrelevant object; wherein the unrelated object and the object are attributed to different recommended information types.

In the above scheme, the semantic analysis module is further configured to extract a text word representing a semantic role in the interaction information to serve as an interaction object of the interaction information; when the interactive object is matched with the object in the video, attitude words corresponding to the interactive object in the interactive information are extracted; and determining the interest degree of the interaction information aiming at the object in the video according to the extracted attitude words.

In the above scheme, the semantic analysis module is further configured to invoke an emotion prediction model to perform the following processing: extracting feature vectors of the attitude words, and mapping the extracted feature vectors into probabilities respectively belonging to different interest degrees; determining the interest degree corresponding to the maximum probability as the interest degree of the interaction information for the object in the video; the emotion prediction model is obtained by training a sample of sample interaction information and the interest degree of the label aiming at the sample interaction information.

In the above solution, the information recommendation apparatus further includes: and the sorting module is used for sorting a plurality of candidate recommendation information pre-associated with the video when the interest degree of the interaction information for the object representation in the video is not determined, and selecting at least one candidate recommendation information to be presented according to the sorting result.

In the above scheme, the ranking module is further configured to select candidate recommendation information to be presented by at least one of the following manners: according to the checking heat degree of each candidate recommendation information, performing descending order arrangement on the multiple candidate recommendation information, and selecting at least one previous candidate recommendation information for presentation; and according to the probability that each candidate recommendation information accords with the user preference, performing descending order arrangement on the plurality of candidate recommendation information, and selecting at least one previous candidate recommendation information for presentation.

In the above scheme, the recommending module is further configured to present the recommending information during the video playing process, and/or present the recommending information when the video playing is finished.

In the above scheme, the recommending module is further configured to jump to a page corresponding to a link carried by the recommending information in response to a trigger operation for the recommending information.

In the above scheme, the recommending module is further configured to present the recommending information when the object appears again in the process of playing a new video, and/or present the recommending information when playing of the new video is finished.

An embodiment of the present invention provides an electronic device, including:

a memory for storing computer executable instructions;

and the processor is used for realizing the information recommendation method provided by the embodiment of the invention when executing the computer executable instructions stored in the memory.

The embodiment of the invention provides a computer-readable storage medium, which stores computer-executable instructions and is used for causing a processor to execute the computer-readable storage medium to realize the information recommendation method provided by the embodiment of the invention.

The embodiment of the invention has the following beneficial effects:

in the video presenting process, the interaction information representing the interest degree of the user for the object in the video is captured, and the recommendation information meeting the interest degree of the user is selected for recommendation.

Drawings

Fig. 1 is a schematic structural diagram of an information recommendation system 100 according to an embodiment of the present invention;

fig. 2 is a schematic structural diagram of an electronic device 500 according to an embodiment of the present invention;

FIG. 3 is a flowchart illustrating an information recommendation method according to an embodiment of the present invention;

fig. 4A and 4B are schematic flowcharts of an information recommendation method according to an embodiment of the present invention;

FIG. 5 is a flowchart illustrating an information recommendation method according to an embodiment of the present invention;

fig. 6A and 6B are schematic diagrams of application scenarios of information recommendation provided by the related art;

fig. 7A and fig. 7B are schematic diagrams of application scenarios provided by an embodiment of the present invention;

FIG. 8 is a flowchart illustrating an information recommendation method according to an embodiment of the present invention;

fig. 9 is a schematic structural diagram of a promotion link according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail with reference to the accompanying drawings, the described embodiments should not be construed as limiting the present invention, and all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict.

Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. The terminology used herein is for the purpose of describing embodiments of the invention only and is not intended to be limiting of the invention.

Before further detailed description of the embodiments of the present invention, terms and expressions mentioned in the embodiments of the present invention are explained, and the terms and expressions mentioned in the embodiments of the present invention are applied to the following explanations.

1) The terminal comprises a client, and an application program which runs in the terminal and is used for providing various services, such as a video client, an instant messaging client, a browser client, an education client or a live broadcast client and the like.

2) In response to the condition or state on which the performed operation depends, one or more of the performed operations may be in real-time or may have a set delay when the dependent condition or state is satisfied; there is no restriction on the order of execution of the operations performed unless otherwise specified.

The embodiment of the invention provides an information recommendation method, an information recommendation device, electronic equipment and a computer-readable storage medium, which can realize accurate recommendation on the basis that a user watches videos. An exemplary application of the electronic device provided by the embodiment of the present invention is described below, and the electronic device provided by the embodiment of the present invention can be implemented as various types of user terminals such as a notebook computer, a tablet computer, a desktop computer, a set-top box, and a mobile device (e.g., a mobile phone, a portable music player, a personal digital assistant, a dedicated messaging device, a portable game device). In the following, an exemplary application will be explained when the device is implemented as a terminal.

Referring to fig. 1, fig. 1 is a schematic structural diagram of an information recommendation system 100 according to an embodiment of the present invention. The information recommendation system 100 includes: the server 200, the network 300, and the terminal 400 will be separately described.

The server 200 is a background server of the client 410, and is configured to respond to a data acquisition request of the client 410 and send a corresponding video to the client 410; and is further configured to receive interaction information sent by the client 410, determine recommendation information according to the interaction information (a specific implementation manner of determining recommendation information according to interaction information will be described in detail below), and send the recommendation information to the client 410.

The network 300, which is used as a medium for communication between the server 200 and the terminal 400, may be a wide area network or a local area network, or a combination of both.

And the terminal 400 is used for operating the client 410. The client 410 is used for receiving the video sent by the server 200 and playing the video in a human-computer interaction interface; the server is further configured to respond to an interactive operation of the user on the video and send interactive information corresponding to the interactive operation to the server 200; and is further configured to receive recommendation information sent by server 200, and present the recommendation information in the human-computer interaction interface.

Here, the client 410 may be an Application (APP) having a video playing function or a live broadcast function, such as a live broadcast APP, a microblog APP, or a short video APP; the browser can also be a browser with a video playing function or a live broadcasting function; but also a video applet or live applet that can be embedded into any APP.

The embodiment of the invention can be realized by means of Cloud Technology (Cloud Technology), which is a hosting Technology for unifying series resources such as hardware, software, network and the like in a wide area network or a local area network to realize the calculation, storage, processing and sharing of data.

The cloud technology is based on the general names of network technology, information technology, integration technology, management platform technology, application technology and the like applied in the 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 technical network systems require a large amount of computing and storage resources, for example, video portals.

As an example, the server 200 may be an independent physical server, may be a server cluster or a distributed system formed by a plurality of physical servers, and may also be a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a web service, cloud communication, a middleware service, a domain name service, a security service, a CDN, and a big data and artificial intelligence platform. The terminal 400 may be, but is not limited to, a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart television, a smart watch, and the like. The terminal 400 and the server 200 may be directly or indirectly connected through wired or wireless communication, and the embodiment of the present invention is not limited thereto.

The embodiment of the invention can be applied to various video watching scenes, such as a live shopping scene, a short video watching scene and the like. Taking a live shopping scene as an example, in the process of watching live through the client 410, a user sends a barrage and/or comments for commodities recommended by a main broadcast in live broadcast; the client 410 sends the barrage and/or the comment to the server 200, the server 200 determines the interest degree of the user for the products recommended by the anchor according to the barrage and/or the comment, determines recommendation information (such as a purchase link) according to the interest degree, and sends the recommendation information to the client 410; the client 410 presents the recommendation information to the human-computer interaction interface for the user to trigger to view.

Next, a structure of an electronic device according to an embodiment of the present invention is described, where the electronic device may be the terminal 400 shown in fig. 1, referring to fig. 2, fig. 2 is a schematic structural diagram of an electronic device 500 according to an embodiment of the present invention, and the electronic device 500 shown in fig. 2 includes: at least one processor 510, memory 550, at least one network interface 520, and a user interface 530. The various components in the electronic device 500 are coupled together by a bus system 540. It is understood that the bus system 540 is used to enable communications among the components. The bus system 540 includes a power bus, a control bus, and a status signal bus in addition to a data bus. For clarity of illustration, however, the various buses are labeled as bus system 540 in fig. 2.

The Processor 510 may be an integrated circuit chip having Signal processing capabilities, such as a general purpose Processor, a Digital Signal Processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware components, or the like, wherein the general purpose Processor may be a microprocessor or any conventional Processor, or the like.

The user interface 530 includes one or more output devices 531 enabling presentation of media content, including one or more speakers and/or one or more visual display screens. The user interface 530 also includes one or more input devices 532, including user interface components to facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

The memory 550 may be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid state memory, hard disk drives, optical disk drives, and the like. Memory 550 optionally includes one or more storage devices physically located remote from processor 510.

The memory 550 may comprise volatile memory or nonvolatile memory, and may also comprise both volatile and nonvolatile memory. The nonvolatile Memory may be a Read Only Memory (ROM), and the volatile Memory may be a Random Access Memory (RAM). The memory 550 described in connection with embodiments of the invention is intended to comprise any suitable type of memory.

In some embodiments, memory 550 can store data to support various operations, examples of which include programs, modules, and data structures, or subsets or supersets thereof, as exemplified below.

An operating system 551 including system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, etc., for implementing various basic services and processing hardware-based tasks;

a network communication module 552 for communicating to other computing devices via one or more (wired or wireless) network interfaces 520, exemplary network interfaces 520 including: bluetooth, wireless compatibility authentication (WiFi), and Universal Serial Bus (USB), etc.;

a presentation module 553 for enabling presentation of information (e.g., a user interface for operating peripherals and displaying content and information) via one or more output devices 531 (e.g., a display screen, speakers, etc.) associated with the user interface 530;

an input processing module 554 to detect one or more user inputs or interactions from one of the one or more input devices 532 and to translate the detected inputs or interactions.

In some embodiments, the information recommendation apparatus provided in the embodiments of the present invention may be implemented in software, and fig. 2 shows an information recommendation apparatus 555 stored in a memory 550, which may be software in the form of programs and plug-ins, and includes the following software modules: interactive presentation module 5551, semantic analysis module 5552 and recommendation module 5553, which are logical and therefore can be arbitrarily combined or further split depending on the functionality implemented. The functions of the respective modules will be explained below.

The information recommendation method provided by the embodiment of the present invention may be executed by the terminal 400 in fig. 1 alone, or may be executed by the terminal 400 and the server 200 in fig. 1 in a cooperation manner.

It should be noted that, the following description is provided by taking an example where the terminal executes the information recommendation method provided by the embodiments of the present invention, and may be implemented by running various forms of computer programs or computer program products, such as the operating system 551, the client 410, the software modules, and the scripts described above, for example.

The following description will be given taking an example in which the terminal 400 in fig. 1 performs the information recommendation method according to the embodiment of the present invention alone. Referring to fig. 3, fig. 3 is a flowchart illustrating an information recommendation method according to an embodiment of the present invention, and will be described with reference to the steps shown in fig. 3.

In step S101, the terminal acquires a video and plays the video.

In some embodiments, the terminal sends a video acquisition request to the server in response to a video viewing operation; the server responds to the video acquisition request and sends a corresponding video to the terminal; and the terminal plays the received video in the video playing interface.

In step S102, the terminal presents the interaction information for the video during the video playing process.

Here, the interaction information can represent the interest degree of the user for the video or the object in the video; the interaction information includes at least one of: comments (information); barrage (information); like (information); back stepping (information); forwarding (information); gifts (information).

In some embodiments, the terminal presents the interactive information for the video in response to the interactive operation for the video during the process of presenting the video.

As an example, the terminal presents a video interaction interface; and responding to the interactive operation aiming at the video received in the video interactive interface, and presenting the interactive information corresponding to the interactive operation.

Here, the video interactive interface may be embedded in a video playing interface, and is suitable for presenting interactive information in the form of barrage, virtual gifts (for example, flowers, smelly eggs, or airplanes), or virtual coins, and the video interactive interface is used for synchronizing the interest level of the user in the object expression in the video.

Taking the example that the interactive information is the barrage, when the video is played on the video playing interface of the terminal, the corresponding barrage is presented on the video playing interface in the presentation form of animation special effects (for example, special effects such as rolling or flying out).

Here, the video interaction interface and the video playing interface may be two independent interfaces, for example, the video interaction interface is located below the video playing interface and is suitable for displaying comment information in the form of text, audio, or animation in an information stream manner; and the method is also suitable for displaying interactive information in the forms of praise, back-stepping and forwarding in a button mode.

Taking the example that the interactive information is comment information, the information streams presented in the video interactive interface have a specific sorting mode, and the sorting mode of the information streams can be sorting according to the posting time of the comment information; the comment information can also be sorted according to the popularity degree of the comment information, for example, the comment information is more frequently viewed by users in the whole network, the popularity degree of the comment information is represented to be higher, and the sorting priority is higher; the ranking may also be performed according to the social distance between the comment information and the user, for example, the more the comment information has been mutually liked, the closer the social distance between the comment information and the user is represented, the higher the ranking priority.

Taking the example that the interactive information is praise, a video interactive interface is presented below a video playing interface of the terminal, wherein the video interactive interface comprises a praise button; when the user carries out the praise operation on the video or the object in the video, the praise button in the video interactive interface is presented in a presentation form different from that before the praise operation, for example, the praise button changes color or shape and the like. It should be noted that the video interaction interface may further include a forward button and/or a back button.

According to the embodiment of the invention, the interaction information is obtained, and the recommendation information which accords with the preference of the user can be accurately selected subsequently according to the interaction information, so that accurate recommendation of the information is realized; and by presenting the interactive information, the attitude of the user in the whole network aiming at the video can be sensed in real time in the process of watching the video, so that the participation of the user in watching the video can be improved.

In step S103, the terminal determines recommendation information according to the interest degree of the object in the video represented by the interaction information.

Here, the terminal may call a corresponding service (e.g., a recommendation service) of the operating system 551, completing the determination process of the recommendation information. The terminal may also invoke a corresponding service (e.g., a recommendation service) of the server, and the determination process of the recommendation information is completed through the server.

Thus, alternative steps of step S103 may be: and the server determines recommendation information according to the interest degree of the object in the video represented by the interaction information and sends the recommendation information to the terminal.

The following describes an example of the determination process of the server completion recommendation information.

In some embodiments, the server selects recommendation information from a plurality of candidate recommendation information according to the interest degree of the object in the video represented by the interaction information, and sends the selected recommendation information to the terminal.

Here, the candidate recommendation information may be uploaded to the server by a user (e.g., a merchant) having a recommendation demand for information; or the information can be obtained by the server through network crawling at a fixed frequency (for example, every 3 days or every 6 hours), so that information recommendation can be realized without uploading recommendation information by a user.

Specific implementations of determining recommendation information are described below.

In some embodiments, the server determines, as the recommendation information, recommendation information of a candidate whose degree of correlation with the object in the video satisfies the correlation condition, among the plurality of candidate recommendation information.

Here, the degree of correlation of the recommendation information characterizes the correlation between the object recommended by the recommendation information and the object in the video. The correlation degree and the interest degree have positive correlation.

Here, when the degree of interest of the interactive information for the object in the video is of interest, the related condition may be: and the recommendation information which is previous in the descending order and has the quantity of the interested quantity threshold value, or the recommendation information of which the correlation degree is greater than the correlation degree threshold value.

The descending order is obtained by ordering the recommendation information of a plurality of candidates according to the degree of correlation from high to low. The quantity of interest threshold may be a default value; or may be a value determined according to the number of candidate recommendation information, wherein the threshold value of the number of interests is proportional to the number of candidate recommendation information. The correlation threshold may be a default value; the degree of correlation may also be determined according to the degree of correlation of the candidate recommendation information, for example, the threshold value of the degree of correlation may be an average value of the degrees of correlation of all the candidate recommendation information.

Here, when the degree of interest of the interactive information for the object in the video is not of interest, the related condition may be: recommending information which is in front in the ascending order and the quantity of which is a non-interesting quantity threshold value, or recommending information of which the correlation degree is not more than the correlation degree threshold value.

The ascending ranking is obtained by ranking the plurality of candidate recommendation information according to the degree of correlation from low to high. The uninteresting quantity threshold may be a default value; or may be a value determined according to the number of candidate recommendation information, wherein the uninteresting number threshold is proportional to the number of candidate recommendation information.

Taking the candidate recommendation information uploaded by the merchant as an example, the server supports the merchant to set the candidate recommendation information and the corresponding correlation degree for the videos or any videos issued by the merchant; or the server only supports the merchant to set candidate recommendation information for videos issued by the merchant or any videos, and the server determines the corresponding correlation degree according to the candidate recommendation information set by the merchant.

The business may set candidate recommendation information for each video, or may set candidate recommendation information for a plurality of videos to be presented in the terminal collectively.

As an example, the process of the server determining the degree of correlation of the candidate recommendation information may be: the server determines an object recommended by the candidate recommendation information; respectively extracting feature vectors of an object recommended by the candidate recommendation information and an object in the video; determining a distance (for example, an Euclidean distance or a cosine distance) between a feature vector of an object recommended by the candidate recommendation information and a feature vector of an object in the video; and determining the correlation degree of the candidate recommendation information according to the distance.

Here, the distance is inversely proportional to the degree of correlation of the candidate recommendation information, and for example, the degree of correlation of the candidate recommendation information may be the inverse number or the opposite number of the distance.

As an example, when the interaction information is interested in the interest degree of the object in the video, the server selects at least one piece of recommendation information in a descending order from a plurality of candidate recommendation information; the descending order is obtained by ordering the recommendation information of a plurality of candidates according to the degree of correlation from high to low.

As another example, when the interest degree of the interaction information for the object in the video is uninteresting, the server selects at least one piece of recommendation information in the ascending order; the ascending ranking is obtained by ranking the plurality of candidate recommendation information according to the degree of correlation from low to high.

In the embodiment of the invention, when a user is interested in an object in a video, recommendation information corresponding to the object with high similarity to the object in the video is preferentially recommended; when the user is not interested in the object in the video, the recommendation information corresponding to the object with low similarity to the object in the video is preferentially recommended, so that the information which is interested by the user can be displayed firstly, the user can see the content which is interested by the user firstly, and the operation path of the user is reduced.

Here, the type of the degree of interest is not limited to two types, i.e., "interested" and "uninteresting", but may also be a type corresponding to a value interval of different degrees of interest, for example, the type corresponding to the value interval [0, 0.3) of the degree of interest is "uninteresting"; the type corresponding to the value interval [0.3, 0.6] of the interest degree is 'normal'; the type corresponding to the value interval (0.6, 1) of the interest degree is 'interest'.

Taking the types of the degrees of interest including "not interested", "general", and "interested" as an example, the server divides the recommendation information of the plurality of candidates into three candidate sets respectively corresponding to the types of the three degrees of interest according to the degree of correlation, for example, the candidate set to which the recommendation information of the candidate whose degree of correlation is in the value interval [0, 0.3) belongs corresponds to "not interested"; the candidate set to which the recommendation information of the candidate with the correlation degree in the value interval [0.3, 0.6] belongs corresponds to "general"; the method comprises the steps of obtaining recommendation information of a candidate in a value interval (0.6, 1) of a relevance degree, wherein the recommendation information of the candidate belongs to a candidate set corresponding to 'interest', sorting the three candidate sets in a descending order according to the relevance degree to obtain three candidate arrangements corresponding to the type of the interest degree, selecting at least one piece of recommendation information which is in front in the descending order corresponding to the interest degree when interaction information aims at the interest degree of an object in a video, so that the recommendation information corresponding to the object with low similarity to the object in the video can be preferentially recommended, selecting at least one piece of recommendation information which is in front in the general descending order when the interest degree of the interaction information aims at the object in the video is general, so that the recommendation information corresponding to the object with the middle similarity to the object in the video can be preferentially recommended, when the interest degree of the interaction information aiming at the object in the video is interested, the server selects at least one piece of recommendation information corresponding to the previous object in the interested descending order, so that the recommendation information corresponding to the object with high similarity to the object in the video can be preferentially recommended.

In the embodiment of the invention, the value interval of the interest degree and the value interval of the correlation degree are matched, so that the workload of the terminal or the server for sorting the candidate recommendation information can be reduced on the basis of realizing accurate recommendation, and the computing resources of the terminal or the server are reduced.

In other embodiments, the server determines, as the recommendation information, recommendation information of a candidate in the recommendation information type corresponding to the degree of interest among the plurality of candidate recommendation information.

Here, the different interest degrees correspond to different recommendation information types, and the interest degrees and the correlation degrees have a positive correlation, and the correlation degree represents the correlation between the object recommended by the recommendation information in the recommendation information type and the object in the video.

Taking the candidate recommendation information uploaded by the merchant as an example, the server supports the merchant to set the candidate recommendation information and the corresponding recommendation information type for the videos or any videos issued by the merchant; or the server only supports the merchant to set candidate recommendation information for videos or any videos issued by the merchant, and the server determines the corresponding recommendation information type according to the candidate recommendation information set by the merchant.

As an example, the process of the server determining the recommendation information type of the candidate recommendation information may be: the server determines an object recommended by the candidate recommendation information; respectively extracting feature vectors of an object recommended by the candidate recommendation information and an object in the video; determining a distance (for example, an Euclidean distance or a cosine distance) between a feature vector of an object recommended by the candidate recommendation information and a feature vector of an object in the video; and determining the recommendation information type of the candidate recommendation information according to the distance.

As an example, when the interaction information is interested in the interest degree of the object in the video, the server selects at least one piece of recommendation information corresponding to the object from the plurality of candidate recommendation information, and/or selects at least one piece of recommendation information corresponding to the related object; wherein the related object and the object belong to the same recommendation information type.

As another example, when the interest degree of the interaction information for the object in the video is uninteresting, the server selects at least one piece of recommendation information corresponding to the irrelevant object from the plurality of candidate recommendation information; wherein the irrelevant objects and the objects belong to different recommended information types.

According to the embodiment of the invention, the information which is interested by the user can be displayed firstly, the user can see the content which is interested by the user firstly, and the operation path of the user is reduced.

Taking the types of the interest degrees including "not interested", "general", and "interested" as an example, the server divides the plurality of candidate recommendation information into three recommendation information types according to the correlation degree, for example, the candidate recommendation information with the correlation degree in the value interval [0, 0.3 ] is a "mutual exclusion type"; the candidate recommendation information with the correlation degree in the value interval [0.3, 0.6] is of other types; the recommendation information of the candidate with the correlation degree in the value range (0.6, 1) belongs to the candidate set corresponding to the same type (when the correlation degree of the recommendation information is 1, the corresponding object is the object in the video), when the interest degree of the interactive information aiming at the object in the video is interested, at least one recommendation information is selected from the recommendation information of the candidate with the recommendation information type of the same type, so that the recommendation information corresponding to the object with high similarity to the object in the video can be preferentially recommended, when the interest degree of the interactive information aiming at the object in the video is general, at least one recommendation information is selected from the recommendation information of the candidate with the recommendation information type of the other type, so that the recommendation information corresponding to the object with medium similarity to the object in the video can be preferentially recommended, when the interest degree of the interactive information aiming at the object in the video is uninteresting, at least one piece of recommendation information is selected from the candidate recommendation information of which the recommendation information type is the 'mutual exclusion type', so that the recommendation information corresponding to the object with low similarity to the object in the video can be preferentially recommended.

It should be noted that the process of determining that the terminal completes the recommendation information is similar to the process of determining that the server completes the recommendation information, and details are not repeated.

In the embodiment of the invention, the interest degree is matched with the recommendation information type of the recommendation information, so that the workload of selecting candidate recommendation information by the terminal or the server can be reduced on the basis of realizing accurate recommendation, and the computing resources of the terminal or the server are reduced.

In step S104, the terminal presents the recommendation information.

Here, the recommendation information presented by the terminal may be recommendation information determined by the terminal or the server in any of the embodiments described above.

In some embodiments, the terminal presents the recommendation information during the video playing process.

As an example, in the video playing process, the terminal acquires recommendation information sent by the server in real time and presents the recommendation information in a recommendation interface. Therefore, recommendation information can be presented in real time according to the attitude of the user aiming at the object in the video, and the real-time performance of recommendation is improved.

Here, the recommendation interface may be embedded in a video playing interface to present recommendation information acquired in real time simultaneously in the process of playing the video; the recommendation interface and the video playing interface can be two independent interfaces, for example, the recommendation interface is located below the video playing interface, so that the recommendation information can be presented without blocking the video.

In some embodiments, the terminal presents the recommendation information when the video playback is finished.

As an example, when the video playing is finished, the terminal acquires the recommendation information sent by the server, and presents the recommendation information in the video playing finishing interface.

In some embodiments, after step S104, the method may further include: and when the object appears again in the process of playing the new video, the terminal presents the recommendation information. Or when the new video playing is finished, the terminal presents the recommendation information.

Here, the new video may be a video that is the same type as or similar to a video that has already been played (i.e., a video that has presented recommendation information); the objects contained in the new video may be the same or similar to the objects contained in the already played video. Because the recommendation information corresponding to the similar videos is more likely to be similar, the recommendation information corresponding to the played video is presented while the new video is presented or after the new video is presented, and the accuracy of information recommendation can be ensured on the basis of reducing the recommendation calculation times.

Here, the new video may be a video played following a video already played; or the time interval between the video and the video which is played is less than the time threshold; and the video playing quantity between the video playing quantity and the video already played is less than the quantity threshold value. Because videos continuously watched by a user often have relevance, the probability that recommendation information corresponding to the relevant videos is similar is high, and thus, the recommendation information corresponding to the played videos is presented while a new video is presented or after the new video is presented, and the accuracy of information recommendation can be ensured on the basis of reducing the recommendation calculation times.

In some embodiments, the recommendation information may include a link for displaying a page corresponding to the link, e.g., a landing advertisement page, a detail page, or a purchase page of goods, etc., when the recommendation information is triggered. When the recommendation information includes a link, after step S104, the method further includes: and the terminal responds to the trigger operation aiming at the recommendation information and jumps to a page corresponding to the link carried by the recommendation information. Here, the page may be an advertisement landing page, a detail page, or a goods purchase page, etc. Therefore, the user can quickly acquire the interested content on the basis of reducing the user operation.

Referring to fig. 4A, fig. 4A is a schematic flowchart of an information recommendation method according to an embodiment of the present invention, and based on fig. 3, step S105 may be included before step S103.

In step S105, the terminal performs semantic analysis processing on the interaction information to obtain a degree of interest of the object in the video represented by the interaction information.

In some embodiments, the terminal may call a corresponding service (e.g., a semantic analysis service) of the operating system 551, and perform a semantic analysis process on the interactive information. The terminal can also call a corresponding service (e.g., a semantic analysis service) of the server, and the semantic analysis processing process of the interactive information is completed through the server.

Thus, alternative steps of step S105 may be: and the server performs semantic analysis processing on the interactive information to obtain the interest degree of the interactive information representation aiming at the object in the video.

As an example, when the execution subject of step S103 is a server, alternative steps of step S105 may be: the terminal carries out semantic analysis processing on the interactive information to obtain an analysis result of the interactive information representation aiming at the interest degree of the object in the video; and the terminal sends the analysis result to the server so that the server can determine the recommendation information according to the analysis result.

In the following, taking the semantic analysis processing process of the interactive information completed by the server as an example, a specific implementation of the semantic analysis will be described.

In some embodiments, the server extracts the text words in the interactive information, which characterize semantic roles, to serve as interactive objects of the interactive information; when the interactive object is matched with the object in the video, attitude words corresponding to the interactive object in the interactive information are extracted; and determining the interest degree of the interactive information aiming at the object in the video according to the extracted attitude words.

In some embodiments, text words that characterize semantic characters may not be included in the interactive information, for example, the comment is "very like" and "very good", so that it is uncertain whether the interactive object in the interactive information and the object in the video match through semantic recognition, and thus the interest level of the interactive information in the object in the video cannot be determined.

In this regard, in some embodiments, the server determines a publication time of the interactive information and determines a current object presented in the video corresponding to the publication time; when a current object presented in the video is matched with an object in the video, attitude words corresponding to the interactive object in the interactive information are extracted; and determining the interest degree of the interactive information aiming at the object in the video according to the extracted attitude words. Therefore, the object preferred by the user can be accurately known by carrying out object identification on the video synchronization, so that the recommendation accuracy is further improved.

As an example, the process of determining, by the server, the interest level of the interaction information for the object in the video according to the extracted attitude word may be: determining the label of the extracted attitude word through an interest word bank; the label is used for indicating the interest degree of the corresponding interaction information for the object in the video.

Taking the types of the degree of interest as "interesting" and "uninteresting", the word bank of interest includes a word bank of interest and a word bank of uninteresting, for example. Words with positive attitudes like "or" very good "are included in the word bank of interest, and words with negative attitudes like" dislike "or" bad "are included in the word bank of no interest.

When the extracted attitude word label is an interested word bank, the corresponding interaction information is interested in the interested degree of the object in the video; and when the label of the extracted attitude word is a non-interest word bank, the interest degree of the corresponding interaction information aiming at the object in the video is not interested.

As an example, the process of the server determining the interaction object of the interaction information may be: the server calls the named entity recognition model to execute the following processing: extracting character feature vectors corresponding to each character in the interactive information; determining a label corresponding to each character in the interactive information according to the character feature vector corresponding to each character in the interactive information; and determining an entity of the interactive information according to the label corresponding to each character, and taking the entity as an interactive object of the interactive information. Therefore, the interactive object of the interactive information can be accurately identified in a machine learning mode, so that the interest degree of the user on the object in the video can be accurately judged, and the accuracy of information recommendation is improved.

Here, the named entity recognition model is obtained by training sample interaction information and an object labeled to the sample interaction information as a sample.

As an example, the process of determining, by the server, the interest level of the interaction information for the object in the video according to the extracted attitude word may be: the server calls the emotion prediction model to execute the following processing: extracting feature vectors of the attitude words, and mapping the extracted feature vectors into probabilities respectively belonging to different interest degrees; and determining the interest degree corresponding to the maximum probability as the interest degree of the interaction information for the object in the video. Therefore, the interest degree of the interactive information aiming at the object in the video can be accurately determined in a machine learning mode, so that the interest degree of the user on the object in the video can be accurately judged, and the accuracy of information recommendation is improved.

Here, the emotion prediction model is obtained by training a sample of sample interaction information and a degree of interest of a label for the sample interaction information.

It should be noted that the process of completing the semantic analysis processing of the interactive information by the terminal is similar to the process of completing the semantic analysis processing of the interactive information by the server, and will not be described again.

Referring to fig. 4B, fig. 4B is a schematic flowchart of an information recommendation method according to an embodiment of the present invention, and based on fig. 3, after step S102, step S106 may be included, where step S106 and step S103 may coexist, and step S106 may also replace step S103.

In step S106, when the interest degree of the interaction information for the object representation in the video is not determined, the terminal sorts a plurality of candidate recommendation information pre-associated with the video, and selects at least one candidate recommendation information according to the sorting result.

In some embodiments, the terminal may invoke a corresponding service (e.g., a sort selection service) of the operating system 551 to complete the sort selection process of the candidate recommendation information. The terminal can also call a corresponding service (for example, a ranking selection service) of the server, and the ranking selection process of the candidate recommendation information is completed through the server.

Thus, alternative steps to step S106 may be: when the interest degree of the interaction information for the object representation in the video is not determined, the server sorts a plurality of candidate recommendation information pre-associated with the video, selects at least one candidate recommendation information according to the sorting result, and sends the selected recommendation information to the terminal.

The following describes a specific implementation manner of ranking selection by taking an example in which a server completes a ranking selection process of candidate recommendation information.

In some embodiments, the server arranges the plurality of candidate recommendation information in a descending order according to the viewing heat of each candidate recommendation information, and selects at least one previous candidate recommendation information for presentation.

Taking the commodity recommendation as an example, the server performs descending order arrangement on the plurality of candidate recommendation information according to the sales volume of the commodity for each candidate recommendation information, and selects at least one previous candidate recommendation information for presentation, so that the hottest commodity can be recommended to the user.

In other embodiments, the server performs descending order on the plurality of candidate recommendation information according to the probability that each candidate recommendation information meets the user preference, and selects at least one previous candidate recommendation information for presentation.

As an example, the server invokes the neural network model to perform the following: extracting the feature vector of each candidate recommendation information, mapping the extracted feature vector into the probability according with the preference of the user, performing descending order arrangement on the plurality of candidate recommendation information according to the probability, and selecting at least one previous candidate recommendation information for presentation.

Here, the neural network model is trained using the sample recommendation information and the degree of interest of the label for the sample recommendation information as a sample.

It should be noted that the process of finishing the sorting and selecting of the candidate recommendation information by the terminal is similar to the process of finishing the sorting and selecting of the candidate recommendation information by the server, and will not be described again.

The embodiment of the invention supports the multi-dimensional determination of the recommendation information meeting the user requirements when the interest degree of the object in the video cannot be analyzed according to the interaction information, so as to ensure the accuracy of information recommendation.

Next, an information recommendation method provided by the embodiment of the present invention cooperatively implemented by the terminal 400 and the server 200 in fig. 1 will be described as an example. Referring to fig. 5, fig. 5 is a flowchart illustrating an information recommendation method according to an embodiment of the present invention, and will be described with reference to the steps shown in fig. 5.

In step S501, the terminal acquires a video and plays the video.

In step S502, the terminal presents the interaction information for the video during the video playing process, and sends the interaction information to the server.

In step S503, the server performs semantic analysis processing on the interaction information to obtain a degree of interest of the interaction information representation for the object in the video.

In step S504, the server determines recommendation information according to the interest degree of the object in the video represented by the interaction information, and sends the recommendation information to the terminal.

In step S505, the terminal presents the recommendation information.

It should be noted that the specific implementation manner in step S501 to step S505 is similar to that in step S101 to step S105, and will not be described again here.

In the embodiment of the invention, the server has strong computing capability and high operation speed compared with the terminal, and the process of determining the recommended information is completed through the server, so that the speed of acquiring the recommended information by the terminal can be improved, and the computing resources of the terminal can be reduced.

In the following, the information recommendation method provided by the embodiment of the present invention is described by taking an example that an application scenario is a recommendation of a commodity (or called a product).

Referring to fig. 6A and 6B, fig. 6A and 6B are schematic diagrams of application scenarios of information recommendation provided by the related art. In fig. 6A, it is common in the related art that a merchant promotes a commodity, and then a corresponding commodity promotion link 601 is given.

For various types of promoted videos, if a user browses the videos, after the user sees the videos and is not interested in the videos, other videos can be selected for watching, and the exposure opportunity loses meaning for merchants. For many long videos, the ad slot setting may exist all the time during the playing of the video, for example, in fig. 6B, a fixed ad slot 602 is always presented in the playing interface of the video. Wherein the ad slots are configured based on the operation, and may or may not be related to the video content. Therefore, the advertisement is difficult to be combined with the user's preference, and thus cannot arouse the user's interest and thus cannot arouse the user's click.

In the related art, the fixed configuration of the ad slots allows the exposure of the video to be attracted to only a portion of users who are interested in the video, whereas the exposure of the video is meaningless for many users who are not interested in the video.

Aiming at the problems, the embodiment of the invention adds a plurality of unfixed advertisement positions in addition to the existing advertisement positions, selects different advertisement positions to display according to the conditions of comments, barracks, praises and the like issued by users at the end of a video, or sequences a plurality of advertisement positions configured by merchants so as to display the advertisement positions interested by the users firstly, so that the users can see the contents interested by the users firstly, and the operation path of the users is reduced.

When a merchant uploads a video, the merchant is allowed to configure a relevant recommended advertisement position to serve as an advertisement link (or a video promotion link, namely the candidate recommendation information) of the video, and when a user watches the video, if the user is interested in the promoted content in the video (or a product promoted in the video, namely the object in the video), the user can jump to a mall page or an advertisement detail page and the like by directly clicking the advertisement link. However, in the case that only one advertisement link is configured for one video, the link cannot play an essential role in diversion for users who dislike the video advertisement, and therefore, the embodiment of the invention allows a merchant to configure a plurality of advertisement links, and advertisement content configured by the advertisement links is allowed to have completely opposite attributes. For example, A and B are two completely different commodities, a merchant is allowed to configure an advertisement link of the commodity B in a video advertisement for promoting the commodity A, and when a user finishes watching a video, a client analyzes the preference of the user for the commodity according to a transmitted barrage and a transmitted video comment of the user in the watching process, and gives a corresponding advertisement link based on an analysis result, so that the possibility of clicking by the user is increased.

The following description will be made with respect to a client (or simply referred to as a merchant) on the merchant side and a client (or simply referred to as a user) on the user (i.e., consumer) side, respectively.

Concrete implementation mode for merchant

Referring to fig. 7A, fig. 7A is a schematic view of an application scenario provided in the embodiment of the present invention. In fig. 7A, when a merchant prepares to publish a promotional video, creation of the video is performed first, after the creation of the video is completed, a configuration button 701 is presented in an interface for publishing the video, and when the merchant triggers the configuration button 701, a link configuration interface is presented. Options are provided in the link configuration interface for the merchant to configure video promotional links, titles of videos, cover art, and the like. For the configuration of the promotion links, the embodiment of the invention allows a plurality of promotion links to be configured for one promotion video, and the commodities in the webpage of the promotion links can be commodities promoted in the video (or called video commodities and video promotion commodities) or other commodities different from the commodities promoted in the video.

The goods in the linked webpage can be the promoted content (or called object) in the video, and also can be other goods of the same type as the video promoted goods.

Here, the merchant needs to fill in keywords of the promotion video in a link configuration interface, and adds a plurality of commodity link contents according to the needs of the merchant, mainly including promoted links and promoted cover drawings. The type field is used for configuring the relationship between the commodities promoted by the video and the promotion link by the merchant. The same type (or called the same type) in the link configuration interface indicates that the commodities corresponding to the promotion link and the video commodities are of the same type, the mutually exclusive type (or called different type) indicates that the commodities corresponding to the promotion link and the video commodities are in a mutually exclusive relationship, and the other types (or called related types and similar types) indicate that the relevance of the commodities corresponding to the promotion link and the video commodities is low. After the configuration is completed, the merchant can publish the video, and thus, the operation of the client at the merchant side is completed.

(II) specific implementation mode for user

For the user, when the user views the video published by the publisher (i.e., the above-mentioned merchant), the user can normally view the promotion link configured by the merchant, as shown in fig. 6A.

The user can see the promotion link configured by the merchant when watching the video normally, and in the process of watching the video, when the user sends a barrage aiming at the video or comments are made on the video, the embodiment of the invention can carry out semantic analysis on the barrage and the comments sent by the user. Thus, different promotion links are displayed based on the result of semantic analysis at the end of video watching.

When the semantic analysis result shows that the user is interested in the commodities promoted in the merchant video, links of the same type as the video promoted commodities are selected from a plurality of promotion links configured by the merchant for display; when the semantic analysis result shows that the user has no interest in the commodities promoted in the merchant video, a link which is mutually exclusive with the video promoted commodities is selected from a plurality of promotion links configured by the merchant for displaying. And based on the result of semantic analysis, performing personalized sequencing on a plurality of promotion links configured by the merchant, and preferentially displaying the promotion links which are interested by the user. And displaying the sorted promotion links after finishing the front cover after the video is played.

Referring to fig. 7B, fig. 7B is a schematic view of an application scenario provided in the embodiment of the present invention, in fig. 7B, when a result of semantic analysis indicates that a user is interested in a commodity promoted in a merchant video, in an advertisement display area 702, a same type of promotion link is preferentially displayed, and then other types of promotion links are displayed.

Referring to fig. 8, fig. 8 is a schematic flowchart of an information recommendation method provided in an embodiment of the present invention, and a specific implementation of the information recommendation method provided in the embodiment of the present invention will be described with reference to fig. 8.

The implementation manner of the embodiment of the invention is summarized as follows:

firstly, a client acquires a promotion link configured by a merchant, and collects a barrage and comments in a video watching process to a server for semantic analysis; and then the server processes the promotion link according to the semantic analysis result. The client is responsible for collection of the promotion links, video playing and display of the promotion links after playing is finished; and the server is responsible for performing semantic analysis processing on the barrage and the comments.

The embodiment of the invention is realized in a specific way that:

in step S801, the client acquires a video uploaded by a merchant and a promotion link configured for the video, and sends the video and the corresponding promotion link to a server (i.e., a video platform).

In step S802, the server checks the video and the corresponding promotion link, and after the check is passed, the video is distributed to the client.

In some embodiments, the server may perform a violation review on the video and the corresponding promotion link, for example, whether the violation content includes pornography, violence, and the like, and when the violation content is included, the review does not pass.

In step S803, the client obtains a corresponding barrage and/or comment in response to an interactive operation when the user watches the video, and submits the barrage and/or comment to the server.

In step S804, the server performs semantic analysis processing on the barrage and/or the comment to obtain an analysis result.

In step S805, the server returns a promotion link to the client according to the analysis result.

In some embodiments, when the analysis result indicates that the user is not interested in the video promotion commodity, a promotion link of a mutually exclusive type configured by the merchant is returned to the client; when the analysis result represents that the user is interested in the video promotion commodity, the promotion links of the same type configured by the merchant are returned to the client; and when the analysis result cannot judge whether the user is interested in the video promotion commodity, the promotion links configured by the merchant are subjected to personalized sequencing and then returned to the client.

In step S806, the client presents the received promotion link on the human-computer interaction interface at the end of the video.

In some embodiments, the user may trigger the corresponding promotion link according to his/her preference.

The logic implementation of the client and the server in the above process is described in detail below.

Logic implementation of client

The client is responsible for uploading the promoted videos, the promoted links and other contents to the server, and for the types of the promoted links, different promoted links need to be distinguished, for example, the relationship between each promoted link and the videos released by merchants (namely, the relationship between the promoted links and the commodities promoted in the videos). Therefore, in the embodiment of the present invention, three types of promotion links are designed, which are "the same type", "mutually exclusive type", and "other types", respectively.

It should be noted that the types of promotion links herein can be extended, and are not limited to the above three types.

The relationship between the promotion links of the three types and the promotion videos published by the merchants is shown in table 1, and table 1 is a relationship table between the promotion links and the promotion videos published by the merchants.

TABLE 1 relationship Table of promotional links with promotional videos published by merchants

In addition to the above three types of correlation degrees of 0, 0.5, and 1, other generalized types of correlation degrees may also be added. The embodiment of the invention takes three types as examples, and when a merchant finishes making the promotion video, fills in keywords of the video, and finishes configuring the promotion link, the video and configuration information are uploaded through the client. As such, the data sent by the client to the server includes: the promotional video, keywords of the promotional video, and a corresponding plurality of promotional links.

And when the video playing is finished, the client sends the barrage and/or comments issued during the period that the user watches the video to the server, acquires the sorted promotion links from the server, and displays the promotion links in a human-computer interaction interface according to the number and sequence of the returned promotion links.

Here, the constituent elements of each promotion link are as shown in fig. 9, and fig. 9 is a schematic structural diagram of a promotion link provided in an embodiment of the present invention.

In fig. 9, each promotion link is presented in the interface in a style in which a picture (ImageView)901, a text (TextView) 903, and a Button (Button)902 are combined into an Item (Item), and each Item, after being triggered, will load a link configured by the merchant through a web view (WebView) to present specific content of the merchant promotion. When the promotion link returned by the server is more than one, the client side can display more contents by forming a vertical list (RecycleView) by a plurality of items.

Logical implementation of server

When the distance from the end of the video is 5 seconds (here, the duration can be adjusted according to the length of the video), the client sends the barrage and/or comment published during the video watching period of the user to the server, and the data sent by the client to the server also comprises the Identification (ID) of the video. In this way, the server can determine the video currently being watched by the user according to the identification of the video. After receiving the data sent by the client, the server firstly searches the video currently watched by the user from the media asset library according to the identification of the video, and then searches the video keywords configured by the merchant according to the video.

The data required by the server for semantic analysis includes: merchant configured video keywords, and user posted barrages and/or comments. The output result of the voice analysis is the sorted promotion links.

For example, the video keyword configured by the merchant is "android phone" (here, the keyword may be more specifically, for example, the phone model, etc., and the keyword may be configured in plural); the merchant configures three promotion links, namely a link A (namely the promotion link of the mobile phone) of the same type, a link B (the promotion link of the television) of the mutually exclusive type and a link C (the promotion link of another mobile phone) of the other type; the comment made by the user is 'bad for the mobile phone'; the published barrage is "buy the mobile phone separately".

The server obtains comments and barrage of the user, finds out that the keywords configured for the video watched by the user are 'android mobile phones', and the promotion link A, the promotion link B and the promotion link C. The server carries out semantic analysis on the comments and the barrage, the main logic of the semantic analysis is to search semantic roles, for example, the 'one' in 'the mobile phone is not good' refers to a commodity popularized in the video. After the semantic role is analyzed, attitude words of the user for the commodity, such as 'bad' and 'buy other', are determined, and other words representing more attitude positive and negative can be quickly retrieved from the barrage and comments of the user by the server in a machine learning mode. And the server combines the semantic role with the user attitude word to obtain a semantic analysis result. After semantic analysis in this example, the server considers that the user is not interested in the commodities promoted by the merchant video, so that the promotion links with the commodity relevance higher than 50% in the promotion links configured by the merchant are removed, that is, link a is removed, and link B and link C are sorted according to the relevance. The final results returned to the client are link C and link B. And after receiving the link, the client displays the created Item on the page with the end of the video for the user to click.

In some embodiments, the types of promotion links may be added in more categories than the three above, where each category has different relevance (i.e., relevance to the promoted video merchandise), such as "similar merchandise" or "substitute merchandise".

In some embodiments, the client collects the factors for analyzing the semantics, and in addition to the above barrage and comments, praise, forward, or back-step and the like can be added to determine the preferred interactive operation of the user.

In some embodiments, in addition to being able to be presented at the end of the video, the promotional links may also be presented at fixed locations based on different behaviors of the user each time. For example, after the user approves the video, the corresponding promotion link is displayed below the video, and the display sequence of the promotion link is refreshed every time the user makes a comment.

The embodiment of the invention supports a merchant to configure the promotion links of a plurality of commodities, and performs personalized screening on the configured promotion links according to the preferences expressed by the user in the process of watching the promotion video, so that the user can preferentially see the most interesting commodities and the commodities which are not interested by the user are shielded, and more click rate is brought to each video exposure of the merchant.

Continuing with the exemplary structure of the information recommendation device 555 provided by the embodiments of the present invention implemented as software modules, in some embodiments, as shown in fig. 2, the software modules stored in the information recommendation device 555 in the memory 550 may include:

the interactive presentation module 5551 is configured to present interactive information for a video during a video playing process;

the semantic analysis module 5552 is configured to perform semantic analysis processing on the interaction information to obtain a degree of interest, represented by the interaction information, for an object in the video;

the recommendation module 5553 is configured to determine recommendation information according to the interest degree of the object in the video represented by the interaction information; and presenting the recommendation information.

In the above solution, the interactive presentation module 5551 is further configured to present a video interactive interface; and responding to the interactive operation aiming at the video received in the video interactive interface, and presenting the interactive information corresponding to the interactive operation.

In the above solution, the recommending module 5553 is further configured to determine, as the recommendation information, recommendation information of a candidate whose degree of correlation with an object in the video satisfies a correlation condition; wherein the degree of correlation characterizes the correlation between the object recommended by the recommendation information and the object in the video; the degree of correlation has a positive correlation with the degree of interest.

In the above solution, the recommending module 5553 is further configured to select at least one piece of recommendation information in a descending order when the interaction information is interested in the interest degree of the object in the video; the descending order is obtained by sorting the plurality of candidate recommendation information according to the correlation degree from high to low; when the interest degree of the interaction information aiming at the object in the video is uninteresting, selecting at least one piece of previous recommendation information in ascending order; the ascending order is obtained by sorting the plurality of candidate recommendation information according to the correlation degree from low to high.

In the foregoing solution, the recommending module 5553 is further configured to determine candidate recommendation information in the recommendation information types corresponding to the interest degree as the recommendation information; the different interest degrees correspond to the different recommendation information types, and positive correlation exists between the interest degrees and the correlation degrees, and the correlation degrees represent the correlation between objects recommended by the recommendation information in the recommendation information types and the objects in the video.

In the above solution, the recommending module 5553 is further configured to, when the interaction information is interested in the interest degree of an object in the video, select at least one piece of recommendation information corresponding to the object, and/or select at least one piece of recommendation information corresponding to a related object; wherein the related object and the object belong to the same recommendation information type; when the interest degree of the interaction information for the object in the video is uninteresting, selecting at least one piece of recommendation information corresponding to the irrelevant object; wherein the unrelated object and the object are attributed to different recommended information types.

In the above scheme, the semantic analysis module 5552 is further configured to extract a text word representing a semantic role in the interaction information, so as to serve as an interaction object of the interaction information; when the interactive object is matched with the object in the video, attitude words corresponding to the interactive object in the interactive information are extracted; and determining the interest degree of the interaction information aiming at the object in the video according to the extracted attitude words.

In the above solution, the semantic analysis module 5552 is further configured to invoke an emotion prediction model to perform the following processing: extracting feature vectors of the attitude words, and mapping the extracted feature vectors into probabilities respectively belonging to different interest degrees; determining the interest degree corresponding to the maximum probability as the interest degree of the interaction information for the object in the video; the emotion prediction model is obtained by training a sample of sample interaction information and the interest degree of the label aiming at the sample interaction information.

In the above solution, the information recommendation device 555 further includes: and the sorting module is used for sorting a plurality of candidate recommendation information pre-associated with the video when the interest degree of the interaction information for the object representation in the video is not determined, and selecting at least one candidate recommendation information to be presented according to the sorting result.

In the above solution, the recommending module 5553 is further configured to present the recommendation information during the video playing process, and/or present the recommendation information when the video playing is finished.

In the foregoing solution, the recommending module 5553 is further configured to jump to a page corresponding to a link carried by the recommendation information in response to a trigger operation for the recommendation information.

In the above solution, the recommending module 5553 is further configured to present the recommendation information when the object appears again in the process of playing a new video, and/or present the recommendation information when the playing of the new video is finished.

Embodiments of the present invention provide a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to execute the artificial intelligence based problem processing method according to the embodiment of the invention.

Embodiments of the present invention provide a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, cause the processor to execute an information recommendation method provided by an embodiment of the present invention, for example, the information recommendation method shown in fig. 3, fig. 4, fig. 5, or fig. 8, where the computer includes various computing devices including an intelligent terminal and a server.

In some embodiments, the computer-readable storage medium may be memory such as FRAM, ROM, PROM, EP ROM, EEPROM, flash memory, magnetic surface memory, optical disk, or CD-ROM; or may be various devices including one or any combination of the above memories.

In some embodiments, the computer-executable instructions may be in the form of programs, software modules, scripts or code written in any form of programming language, including compiled or interpreted languages, or declarative or procedural languages, and they may be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

By way of example, computer-executable instructions may correspond, but do not necessarily have to correspond, to files in a file system, and may be stored in a portion of a file that holds other programs or data, e.g., in one or more scripts in a hypertext markup language document, in a single file dedicated to the program in question, or in multiple coordinated files (e.g., files that store one or more modules, sub-programs, or portions of code).

By way of example, computer-executable instructions may be deployed to be executed on one computing device or on multiple computing devices at one site or distributed across multiple sites and interconnected by a communication network.

In summary, the embodiments of the present invention have the following beneficial effects:

(1) acquiring the interactive information, and accurately selecting recommendation information according with the user preference according to the interactive information subsequently, so as to realize accurate recommendation of the information; and by presenting the interactive information, the attitude of other users aiming at the video can be sensed in real time in the process of watching the video by the user, so that the participation of the user in watching the video can be improved.

(2) When a user is interested in an object in a video, preferentially recommending recommendation information corresponding to the object with the highest similarity to the object in the video; when the user is not interested in the object in the video, the recommendation information corresponding to the object with the lowest similarity to the object in the video is preferentially recommended, so that the information which is interested by the user can be displayed firstly, the user can see the content which is interested by the user firstly, and the operation path of the user is reduced.

(3) The interest degree is matched with the recommendation information type of the recommendation information, so that the workload of selecting candidate recommendation information by the terminal or the server can be reduced on the basis of realizing accurate recommendation, and the computing resources of the terminal or the server are reduced.

(4) The value intervals of the interesting degrees and the value intervals of the relevant degrees are matched, so that the workload of the terminal or the server for sorting the candidate recommendation information can be reduced on the basis of realizing accurate recommendation, and the computing resources of the terminal or the server are reduced.

(5) And presenting recommendation information corresponding to the played video while presenting the new video or after finishing presenting the new video, so that the accuracy of recommendation can be ensured on the basis of reducing the recommendation calculation times.

The above description is only an example of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, and improvement made within the spirit and scope of the present invention are included in the protection scope of the present invention.

Claims

1. An information recommendation method, characterized in that the method comprises:

presenting a video interaction interface different from a video playing interface in the video playing process;

responding to the interaction operation aiming at the video received in the video interaction interface, and presenting interaction information corresponding to the interaction operation;

performing semantic analysis processing on the interaction information to obtain the interest degree of the interaction information representation aiming at the object in the video, wherein the interest degree comprises at least two types;

acquiring a plurality of candidate recommendation information and the correlation degree between an object recommended by the candidate recommendation information and an object in the video;

wherein the degree of correlation characterizes a correlation between an object recommended by the candidate recommendation information and an object in the video; the correlation degree and the interest degree have positive correlation;

based on the correlation degree, dividing the candidate recommendation information into recommendation information types corresponding to the interest degrees of the types;

determining a target recommendation information type from a plurality of recommendation information types based on the interest degree;

based on the target recommendation information type, selecting at least one piece of recommendation information from candidate recommendation information corresponding to the target recommendation information type;

determining a recommendation position for displaying the recommendation information based on the interest degree;

presenting the selected recommendation information at the recommendation position of a recommendation interface different from the video playing interface.

2. The method of claim 1, further comprising:

when the type of the interest degree of the interactive information aiming at the object in the video is interest, selecting at least one piece of previous recommendation information in descending order;

the descending order is obtained by sorting the plurality of candidate recommendation information according to the correlation degree from high to low;

when the type of the interest degree of the interactive information aiming at the object in the video is uninteresting, selecting at least one piece of previous recommendation information in ascending order;

the ascending order is obtained by sorting the plurality of candidate recommendation information according to the correlation degree from low to high.

3. The method according to claim 1, wherein the selecting at least one piece of recommendation information from the candidate recommendation information corresponding to the target recommendation information type based on the target recommendation information type includes:

when the target recommendation information type is the type corresponding to the interest degree of each type and is the interest, selecting at least one piece of recommendation information corresponding to the object and/or selecting at least one piece of recommendation information corresponding to the related object;

wherein the related object and the object belong to the same recommendation information type;

when the target recommendation information type is the type of interest degree corresponding to each type and is uninteresting, selecting at least one piece of recommendation information corresponding to an irrelevant object;

wherein the unrelated object and the object are attributed to different recommended information types.

4. The method of claim 1, wherein the semantic analysis processing the interaction information to obtain the interest level of the object in the video represented by the interaction information comprises:

extracting text words representing semantic roles in the interactive information to serve as interactive objects of the interactive information;

when the interactive object is matched with the object in the video, attitude words corresponding to the interactive object in the interactive information are extracted;

and determining the interest degree of the interaction information aiming at the object in the video according to the extracted attitude words.

5. The method according to claim 4, wherein the determining the interest level of the interaction information for the object in the video according to the extracted attitude word comprises:

calling the emotion prediction model to execute the following processing:

extracting feature vectors of the attitude words, and mapping the extracted feature vectors into probabilities respectively belonging to different interest degrees;

determining the interest degree corresponding to the maximum probability as the interest degree of the interaction information for the object in the video;

the emotion prediction model is obtained by training a sample of sample interaction information and the interest degree of the label aiming at the sample interaction information.

6. The method of claim 1, further comprising:

when the interest degree of the interaction information for the object representation in the video is not determined, sorting a plurality of candidate recommendation information pre-associated with the video, and sorting the candidate recommendation information

And selecting at least one candidate recommendation information for presentation according to the sorting result.

7. The method of claim 6, wherein the sorting the candidate pieces of recommendation information pre-associated with the video and selecting at least one candidate piece of recommendation information to present according to the sorting result comprises:

selecting candidate recommendation information to be presented by at least one of the following modes:

according to the checking heat degree of each candidate recommendation information, performing descending order arrangement on the multiple candidate recommendation information, and selecting at least one previous candidate recommendation information for presentation;

and according to the probability that each candidate recommendation information accords with the user preference, performing descending order arrangement on the plurality of candidate recommendation information, and selecting at least one previous candidate recommendation information for presentation.

8. The method of claim 1, wherein the presenting the recommendation information comprises:

and presenting the recommendation information in the video playing process, and/or presenting the recommendation information when the video playing is finished.

9. The method of any of claims 1 to 8, wherein after said presenting the recommendation information, the method further comprises:

and when the object appears again in the process of playing the new video, presenting the recommendation information, and/or when the playing of the new video is finished, presenting the recommendation information.

10. An information recommendation apparatus, characterized in that the apparatus comprises:

the interactive presentation module is used for presenting a video interactive interface different from the video playing interface in the video playing process; responding to the interaction operation aiming at the video received in the video interaction interface, and presenting interaction information corresponding to the interaction operation;

the semantic analysis module is used for performing semantic analysis processing on the interaction information to obtain the interest degree of the interaction information representation aiming at the object in the video, and the interest degree comprises at least two types;

the recommendation module is used for acquiring recommendation information of a plurality of candidates and the correlation degree between an object recommended by the recommendation information of the candidates and an object in the video; wherein the degree of correlation characterizes a correlation between an object recommended by the candidate recommendation information and an object in the video; the correlation degree and the interest degree have positive correlation; based on the correlation degree, dividing the candidate recommendation information into recommendation information types corresponding to the interest degrees of the types; determining a target recommendation information type from a plurality of recommendation information types based on the interest degree; based on the target recommendation information type, selecting at least one piece of recommendation information from candidate recommendation information corresponding to the target recommendation information type; determining a recommendation position for displaying the recommendation information based on the interest degree; presenting the selected recommendation information at the recommendation position of a recommendation interface different from the video playing interface.

11. An electronic device, comprising:

a memory for storing computer executable instructions;

a processor for implementing the information recommendation method of any one of claims 1 to 9 when executing the computer-executable instructions stored in the memory.

12. A computer-readable storage medium storing computer-executable instructions for implementing the information recommendation method of any one of claims 1 to 9 when executed by a processor.