CN112104899B

CN112104899B - Information recommendation method and device in live broadcast, electronic equipment and storage medium

Info

Publication number: CN112104899B
Application number: CN202010953065.3A
Authority: CN
Inventors: 陈春勇
Original assignee: Tencent Technology Shenzhen Co Ltd
Current assignee: Tencent Technology Shenzhen Co Ltd
Priority date: 2020-09-11
Filing date: 2020-09-11
Publication date: 2024-03-08
Anticipated expiration: 2040-09-11
Also published as: CN112104899A

Abstract

The application provides an information recommendation method, an information recommendation device, electronic equipment and a computer readable storage medium in live broadcast; the method comprises the following steps: acquiring introduction information of articles to be recommended in a live broadcasting room of a live broadcasting platform; presenting a recommendation page of the live broadcast platform in a human-computer interaction interface; wherein, the recommendation page includes: introduction information of at least one item to be recommended, and a live room portal corresponding to each of the items. According to the method and the device, big data processing based on the cloud technology is used, articles to be recommended in the live broadcasting room can be quickly and intuitively known in the recommendation page of live broadcasting of the user, effective recommendation in live broadcasting can be achieved, and stable operation of live broadcasting in a live broadcasting system is guaranteed.

Description

Information recommendation method and device in live broadcast, electronic equipment and storage medium

Technical Field

The present disclosure relates to internet technology and cloud technology, and in particular, to an information recommendation method and apparatus in live broadcast, an electronic device, and a computer readable storage medium.

Background

Live broadcasting is an important way of information transmission in the Internet, and large data processing based on cloud technology is used for collecting live broadcasting contents of massive anchor and distributing the live broadcasting contents to audiences participating in the anchor in real time. Currently, a host will conduct various recommended activities in a live room. Due to the large number of anchor campaigns that are resident to the live platform, the variety of recommended campaigns involved is also large, and it is often difficult for viewers to select the appropriate live room to view the relevant recommended campaign.

For example, in the related art, a merchant or a store is often used as a recommendation object, so that a user is difficult to intuitively acquire the articles to be recommended in the live broadcast room, and therefore, the user needs to pull a list one by one or enter the live broadcast room to know the articles specifically recommended in the live broadcast room, and invalid recommendation of the articles is often caused, so that unnecessary waste is caused on computing resources and communication resources for recommending the articles in the live broadcast system, and further, the live broadcast stability in the live broadcast platform is affected.

Therefore, for realizing effective recommendation of the articles in the live broadcast, the stable operation of the live broadcast in the live broadcast system is ensured, and no effective solution exists in the related technology.

Disclosure of Invention

The embodiment of the application provides an information recommendation method, an information recommendation device, electronic equipment and a computer readable storage medium in live broadcast, which can realize effective recommendation of articles in live broadcast and ensure stable operation of live broadcast in a live broadcast system.

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

the embodiment of the application provides an information recommendation method in live broadcast, which comprises the following steps:

acquiring introduction information of articles to be recommended in a live broadcasting room of a live broadcasting platform;

presenting a recommendation page of the live broadcast platform in a human-computer interaction interface;

Wherein, the recommendation page includes: the method comprises the steps of recommending at least one item to be recommended, introducing information of the at least one item to be recommended, and a live broadcasting room entrance corresponding to each item.

In the above scheme, the recommendation page further comprises a purchase entry corresponding to each item; the method further comprises the steps of:

in response to a trigger operation of a purchase portal for a target item, displaying a purchase page for purchasing the target item;

wherein the target item is a selected item among the at least one item to be recommended.

In the above scheme, the presenting the recommendation page of the live broadcast platform in the man-machine interaction interface includes:

presenting a live broadcasting room display page of the live broadcasting platform in the man-machine interaction interface;

responding to a recommendation operation received in the live broadcasting room display page, and presenting the recommendation page;

wherein, the recommendation page includes: and introducing information of the articles to be recommended in at least one live broadcasting room displayed in the live broadcasting room display page.

presenting live broadcast content of a live broadcast room in the man-machine interaction interface;

Presenting the recommendation page in response to a recommendation operation received in the live broadcast room;

wherein, the recommendation page includes: and introducing information of the articles to be recommended in the live broadcasting room.

The embodiment of the application provides an information recommendation device in live broadcast, which comprises:

the acquisition module is used for acquiring introduction information of the articles to be recommended in the live broadcasting room of the live broadcasting platform;

the live broadcast presenting module is used for presenting a recommended page of the live broadcast platform in the man-machine interaction interface;

In the above scheme, the obtaining module is further configured to perform real-time voice recognition on live content in at least one live broadcast room in the live broadcast platform to obtain a voice recognition result; and extracting introduction information of a plurality of articles to be recommended from the voice recognition result.

In the above scheme, the obtaining module is further configured to extract audio data from the live broadcast content, and perform frame framing processing on the audio data to obtain a plurality of audio sub-data; extracting acoustic features of each audio sub-data to obtain a plurality of audio vectors; the following is performed for each of the audio vectors: matching the audio vector with each reference template in a reference voice template library to determine the similarity between the audio vector and each reference template, and determining the text information corresponding to the reference template with the highest similarity as the text information corresponding to the audio vector; and combining a plurality of text information corresponding to the audio sub-data one by one to obtain the voice recognition result.

In the above scheme, the acquiring module is further configured to acquire text information in a floating layer included in the live broadcasting room; and in the basic voice template library, determining a plurality of basic templates matched with the text information in the floating layer, and combining the determined basic templates into the reference voice template library.

In the above scheme, the acquiring module is further configured to acquire live broadcast forecast information pre-submitted by at least one live broadcast room in the live broadcast platform; and extracting introduction information of a plurality of articles to be recommended from the live broadcast forecast information.

In the above scheme, the acquiring module is further configured to determine a recommendation index of each item to be recommended according to an item feature in the introduction information; and sorting the plurality of articles to be recommended in a descending order of recommendation indexes, and filtering out introduction information of the articles to be recommended in the later part of the descending order.

In the above scheme, the obtaining module is further configured to filter out introduction information of an article that does not meet the condition of interest from the introduction information of the plurality of articles to be recommended; wherein the condition of interest comprises at least one of: items having an interactive relationship with the live account; the articles with interactive relation with the live account belong to the same type of articles; the main broadcast focused by the live account recommends articles in the live broadcast room; and (3) recommending the articles by the account with social relation with the live account.

In the above scheme, the acquiring module is further configured to acquire a predetermined time of each item to be recommended; filtering out introduction information of the articles to be recommended, wherein the preset time of the introduction information is out of a target time range, from a plurality of articles to be recommended; wherein the type of the target time range includes at least one of: the expected viewing time range of the live account; idle viewing time range of live account numbers.

In the above scheme, the live broadcast presenting module is further configured to display a live broadcast room for introducing the target item in response to a trigger operation of a live broadcast room portal for the target item; wherein the target item is a selected item among the at least one item to be recommended.

In the above scheme, the live broadcast presenting module is further configured to display a live broadcast room for introducing the target object when a difference between the predetermined time of the target object and the trigger operation time does not exceed a time difference threshold; when the difference value between the preset time of the target object and the triggering operation time exceeds a time difference threshold, generating a reminding task, and executing the reminding task before the preset time to present reminding information.

In the above scheme, the live broadcast presenting module is further configured to present, when the man-machine interaction interface presents a live broadcast room for introducing the target object, a purchase entrance of the target object in the live broadcast room in a manner of popup window or floating layer; when the man-machine interaction interface presents a live broadcasting room and the live broadcasting room is not used for introducing the target object, presenting a live broadcasting room entrance of the target object in the live broadcasting room in a popup window or floating layer mode; and when the man-machine interaction interface does not present the live broadcasting room, presenting the live broadcasting room entrance and/or the purchase entrance in a notification message mode.

In the above scheme, the recommendation page further comprises a purchase entry corresponding to each item; the live broadcast presentation module is further used for responding to the triggering operation of the purchase entrance for the target object and displaying a purchase page for purchasing the target object; wherein the target item is a selected item among the at least one item to be recommended.

In the above scheme, the live broadcast presenting module is further configured to present a live broadcast room display page of the live broadcast platform in the man-machine interaction interface; responding to a recommendation operation received in the live broadcasting room display page, and presenting the recommendation page; wherein, the recommendation page includes: and introducing information of the articles to be recommended in at least one live broadcasting room displayed in the live broadcasting room display page.

In the above scheme, the live broadcast presenting module is further configured to present live broadcast content in a live broadcast room in the man-machine interaction interface; presenting the recommendation page in response to a recommendation operation received in the live broadcast room; wherein, the recommendation page includes: and introducing information of the articles to be recommended in the live broadcasting room.

In the above solution, when the recommendation page includes introduction information of a plurality of the articles to be recommended, a manner of ordering the introduction information of the plurality of the articles to be recommended includes at least one of: sorting the introduction information of the plurality of articles to be recommended according to the sequence of the preset time of the articles to be recommended; sorting the introduction information of the plurality of articles to be recommended in a descending order according to the interaction heat of the articles to be recommended; sorting the introduction information of the plurality of articles to be recommended in a descending order according to the preference degree of the articles to be recommended according with the live account; and sorting the introduction information of the plurality of articles to be recommended in a descending order according to the recommendation indexes of the articles to be recommended.

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

A memory for storing computer executable instructions;

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

The embodiment of the application provides a computer readable storage medium, which stores computer executable instructions for realizing the information recommendation method in live broadcast provided by the embodiment of the application when being executed by a processor.

The embodiment of the application has the following beneficial effects:

the method has the advantages that the introduction information of the articles to be recommended in the live broadcasting room can be intuitively obtained before the live broadcasting room is entered, whether the articles are interested in the live broadcasting room or not can be further determined according to the introduction information, the efficiency of human-computer interaction in the recommendation process is improved, the accuracy and the efficiency of recommendation are improved, and the computing resources and the communication resources used when the server sends the recommendation information are saved.

Drawings

Fig. 1A and 1B are schematic views of an application scenario provided by the related art;

fig. 2 is a schematic structural diagram of an information recommendation system 100 in live broadcast according to an embodiment of the present application;

fig. 3 is a schematic structural diagram of an electronic device 500 according to an embodiment of the present application;

fig. 4 is a flowchart of an information recommendation method in live broadcast according to an embodiment of the present application;

Fig. 5 is a flowchart of an information recommendation method in live broadcast according to an embodiment of the present application;

fig. 6 is a flowchart of an information recommendation method in live broadcast according to an embodiment of the present application;

fig. 7 is a flowchart of an information recommendation method in live broadcast according to an embodiment of the present application;

fig. 8 is a flowchart of an information recommendation method in live broadcast according to an embodiment of the present application;

fig. 9A, fig. 9B, fig. 9C, fig. 9D, fig. 9E, and fig. 9F are schematic application scenarios of an information recommendation method in live broadcast provided in an embodiment of the present application;

fig. 10 is a flowchart of an information recommendation method in live broadcast provided in an embodiment of the present application;

FIG. 11 is a schematic flow chart of speech recognition provided in an embodiment of the present application;

fig. 12 is a schematic diagram of speech recognition provided in an embodiment of the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the present application will be described in further detail with reference to the accompanying drawings, and the described embodiments should not be construed as limiting the present application, and all other embodiments obtained by those skilled in the art without making any inventive effort are within the scope of the present application.

In the following description, reference is made to "some embodiments" which describe a subset of all possible embodiments, but it is to be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments and can be combined with one another 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 application belongs. The terminology used herein is for the purpose of describing embodiments of the present application only and is not intended to be limiting of the present application.

Before further describing embodiments of the present application in detail, the terms and expressions that are referred to in the embodiments of the present application are described, and are suitable for the following explanation.

1) In response to a condition or state that is used to represent the condition or state upon which the performed operation depends, the performed operation or operations may be in real-time or with a set delay when the condition or state upon which it depends is satisfied; without being specifically described, there is no limitation in the execution sequence of the plurality of operations performed.

2) A client, an application running in the terminal for providing various services, such as a live client, a video client, or a short video client.

3) The content delivery network (CDN, content Delivery Network), a layer of network architecture added to the internet, is composed of cache servers. The cache servers can all store business contents according to a certain cache policy, when a user accesses a website, the user access is directed to the cache server which works normally and is closest to the website by using a global load technology, and the cache server directly responds to the user request.

With the development of the live broadcast industry of electronic commerce, more and more live broadcast and take goods platforms of electronic commerce are presented. In the related art, referring to fig. 1A and 1B, fig. 1A and 1B are schematic views of application scenarios provided by the related art. In fig. 1A and fig. 1B, the recommendation page takes the live broadcasting room 101 and the store 102 as dimensions, and a user is difficult to intuitively find the articles to be recommended in the live broadcasting, so that the user needs to enter the live broadcasting room or the store one by one to acquire the articles to be recommended, which is not beneficial to intuitively knowing the preferential strength of the articles to be recommended in the live broadcasting, is time-consuming and labor-consuming, and has poor experience.

Aiming at the technical problems, the embodiment of the application provides an information recommendation method in live broadcast, which can realize effective recommendation of articles in live broadcast and ensure stable operation of live broadcast in a live broadcast system. Referring to fig. 2, fig. 2 is a schematic structural diagram of an information recommendation system 100 in live broadcast according to an embodiment of the present application. Wherein, the information recommendation system 100 in live broadcast includes: the server 200, the network 300, and the terminal 400 will be described separately.

The server 200 is a background server of the client 410, and is configured to send, to the client 410, introduction information of an item to be recommended in a living room of the living platform in response to an item recommendation request of the client 410.

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

The terminal 400 is configured to run the client 410, where the client 410 is a client with a live broadcast function. The client 410 is configured to send an item recommendation request to the server 200 in response to a recommendation operation of a user, so as to receive introduction information of items to be recommended in a living room of a living platform sent by the server 200, and present a recommendation page of the living platform in the man-machine interaction interface 411, where the recommendation page includes introduction information of at least one item to be recommended and a living room entry corresponding to each item.

In some embodiments, the terminal 400 implements the information recommendation method in live broadcast provided in the embodiments of the present application by running a computer program, for example, the computer program may be a native program or a software module in an operating system; it may be a local (Native) Application (APP), i.e. a program that needs to be installed in an operating system to run, such as a live APP or a video APP; the method can also be an applet, namely a program which can be run only by being downloaded into a browser environment; but also a live applet or video applet that can be embedded in any APP. In general, the computer programs described above may be any form of application, module or plug-in.

The embodiment of the application can be realized by means of Cloud Technology (Cloud Technology), wherein the Cloud Technology refers to a hosting Technology for integrating serial resources such as hardware, software, network and the like in a wide area network or a local area network to realize calculation, storage, processing and sharing of data.

The cloud technology is a generic term of network technology, information technology, integration technology, management platform technology, application technology and the like based on cloud computing business model application, can form a resource pool, and is flexible and convenient as required. Cloud computing technology will become an important support. Background services of technical network systems require a large amount of computing and storage resources.

As an example, the server 200 may be a stand-alone physical server, a server cluster or a distributed system formed by a plurality of physical servers, or may be a cloud server that provides cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDNs, and basic cloud computing services such as big data and artificial intelligence platforms. The terminal 400 may be, but is not limited to, a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, etc. The terminal 400 and the server 200 may be directly or indirectly connected through wired or wireless communication, which is not limited in the embodiment of the present application.

Next, the structure of the electronic device provided in the embodiment of the present application will be described, where the electronic device may be the terminal 400 shown in fig. 2, referring to fig. 3, and fig. 3 is a schematic structural diagram of the electronic device 500 provided in the embodiment of the present application, and the electronic device 500 shown in fig. 3 includes: at least one processor 510, a memory 550, at least one network interface 520, and a user interface 530. The various components in electronic device 500 are coupled together by bus system 540. It is appreciated that the bus system 540 is used to enable connected communications between these components. The bus system 540 includes a power bus, a control bus, and a status signal bus in addition to the data bus. The various buses are labeled as bus system 540 in fig. 3 for clarity of illustration.

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

The user interface 530 includes one or more output devices 531 that enable presentation of media content, including one or more speakers and/or one or more visual displays. The user interface 530 also includes one or more input devices 532, including user interface components that 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 drives, optical drives, and the like. Memory 550 may optionally include one or more storage devices physically located remote from processor 510.

Memory 550 includes volatile memory or nonvolatile memory, and may also include both volatile and nonvolatile memory. The non-volatile memory may be read only memory (ROM, read Only Me mory) and the volatile memory may be random access memory (RAM, random Access Memor y). The memory 550 described in embodiments herein is intended to comprise any suitable type of memory.

In some embodiments, memory 550 is capable of storing 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 handling 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 handling hardware-based tasks;

network communication module 552 is used to reach other computing devices via one or more (wired or wireless) network interfaces 520, exemplary network interfaces 520 include: bluetooth, wireless compatibility authentication (WiFi), and universal serial bus (USB, universal Serial Bus), etc.;

A presentation module 553 for enabling presentation of information (e.g., a user interface for operating a peripheral device 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;

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

In some embodiments, the in-live information recommendation device provided in the embodiments of the present application may be implemented in a software manner, and fig. 3 shows an in-live information recommendation device 555 stored in a memory 550, which may be software in the form of a computer program, a plug-in, or the like, for example, a video client, a live client, or a short video client. The in-live information recommendation device 555 includes the following software modules: the acquisition module 5551 and live presentation module 5552 are logical and thus may be arbitrarily combined or further split depending on the functionality implemented. The functions of the respective modules will be described hereinafter.

The method for recommending information in live broadcast provided in the embodiment of the present application may be executed by the terminal 400 in fig. 2 alone, or may be executed by the terminal 400 and the server 200 in fig. 2 cooperatively.

Next, an information recommendation method in live broadcast provided in the embodiment of the present application is described as an example by the terminal 400 in fig. 2 alone. Referring to fig. 4, fig. 4 is a flowchart of a method for recommending information in live broadcast according to an embodiment of the present application, and the steps shown in fig. 4 will be described.

It should be noted that the method shown in fig. 4 may be executed by various computer programs executed by the terminal 400, and is not limited to the above-described client 410, such as the operating system 551, the software modules, and the scripts, and therefore the client should not be considered as limiting the embodiments of the present application.

In step S101, introduction information of an item to be recommended in a live room of a live platform is acquired.

Here, the article is a variety of valuable articles (or products), and may be an actual article (e.g., food, clothing), or may be a virtual article (e.g., game props, software) and service (e.g., legal consultation, office replacement, goods dispatch, etc.).

In some embodiments, the client may invoke a corresponding service (e.g., introduction information extraction service) of the terminal, and the process of extracting the introduction information of the item to be recommended is completed through the terminal. The client may also call a corresponding service (e.g., introduction information extraction service) of the server, and the process of extracting the introduction information of the item to be recommended is completed through the server.

As an example, when the client invokes a corresponding service (e.g., introduction information extraction service) of the server to complete the process of extracting the introduction information of the item to be recommended, the replacement step of step S101 is: the method comprises the steps that a server obtains introduction information of articles to be recommended in a live broadcasting room of a live broadcasting platform; and transmitting the introduction information of the articles to be recommended to the client.

In the following, a procedure of extracting introduction information of an item to be recommended by a terminal is described as an example in which a corresponding service (for example, introduction information extraction service) of the terminal is invoked by a client. It should be noted that, the description information of the item to be recommended is extracted by the client calling the corresponding service (for example, description information extraction service) of the server, which is similar to the following description, and will not be repeated.

In some embodiments, introduction information of items to be recommended in all living rooms in the living platform may be obtained; the introduction information of the articles to be recommended in the living broadcast room meeting the interested conditions in the living broadcast platform can also be selectively obtained.

As an example, the condition of interest includes at least one of: a live room having an interactive relationship (e.g., collection, browsing, purchasing, praying or commenting, etc.) with a live account (e.g., a live account logged into a client); live broadcast rooms with interactive relation with live account numbers belong to the same type of live broadcast room; and the live broadcasting room recommended by the account with social relation with the live broadcasting account.

Here, the account number having a social relationship with the live account number may be an account number belonging to the same live platform, or may be an account number of another social platform different from the live platform, such as a microblog account number and a blog account number.

In some embodiments, referring to fig. 5, fig. 5 is a schematic flow diagram of an information recommendation method in live broadcast provided in the embodiments of the present application, and based on fig. 4, step S101 may include step S1011 and step S1012.

In step S1011, real-time speech recognition is performed on live contents of at least one live broadcast room in the live broadcast platform, so as to obtain a speech recognition result.

In some embodiments, audio data is extracted from live content and is subjected to framing processing to obtain a plurality of audio sub-data; extracting acoustic features of each audio sub-data to obtain a plurality of audio vectors; the following processing is performed for each audio vector: matching the audio vector with each reference template in the reference voice template library to determine the similarity between the audio vector and each reference template, and determining the text information corresponding to the reference template with the highest similarity as the text information corresponding to the audio vector; and combining a plurality of text information corresponding to the plurality of audio sub-data one by one to obtain a voice recognition result.

As an example, the obtaining manner of the reference voice template library may include: acquiring character information in a floating layer included in a live broadcasting room; and in the basic voice template library, determining a plurality of basic templates matched with the text information in the floating layer, and combining the determined basic templates into a reference voice template library.

Here, the base speech template library includes audio vectors corresponding to the speech of all words in the base vocabulary.

For example, when the text information in the floating layer is "8-point second cake, only 10-element" is needed, the audio vectors corresponding to the keywords "8-point", "second cake", and "10-element" are searched in the base voice template library, so as to be combined into the reference voice template library, that is, the reference voice template library contains the audio vectors corresponding to the keywords "8-point", "second cake", and "10-element" respectively.

Therefore, the similarity comparison is carried out on each audio vector contained in the audio data extracted from the live broadcast content and the reference templates in the reference voice template library in sequence, so that a voice recognition result is obtained, and compared with the method of directly searching corresponding text information in the basic voice template library, the voice recognition speed is faster, the accuracy is higher, and the consumption of computing resources can be reduced.

In step S1012, introduction information of a plurality of articles to be recommended is extracted from the voice recognition result.

Here, the introduction information includes information for introducing various attributes such as a name, price, predetermined time, and quantity of the item.

For example, when the voice recognition result is "8-point seconds for killing a cake, as long as 10 yuan, very good eating", the introduction information of the item to be recommended includes "8 points", "seconds for killing", "cake" and "10 yuan".

According to the method and the device for recommending the articles, the user can acquire the introduction information of the articles to be recommended, which are broadcast by the anchor in the live broadcasting room, in real time through voice recognition, timeliness of the extracted articles to be recommended can be guaranteed, and therefore accuracy and efficiency of recommendation can be improved.

In some embodiments, referring to fig. 6, fig. 6 is a flowchart of an information recommendation method in live broadcast provided in the embodiment of the present application, and based on fig. 4, step S1013 and step S1014 may be included in step S101.

In step S1013, live preview information submitted in advance by at least one live room in the live platform is acquired.

In some embodiments, the live preview information is determined before the live broadcast starts, and includes information about objects to be introduced by the host, such as a start time, an item name, purchase information, etc., which may be synchronized by the host to the live broadcast platform before the live broadcast starts, or may be extracted from subscription data of the host and the advertisement system. As an example, the presentation form of the live preview information may be in the form of a picture, audio, video, or text, and a process of extracting introduction information of an item to be recommended will be described below with respect to the above forms, respectively.

In step S1014, introduction information of a plurality of items to be recommended is extracted from the live preview information.

Taking the presentation form of the live broadcast forecast information as an example, carrying out image recognition on the live broadcast forecast information to obtain an image recognition result; and extracting introduction information of a plurality of candidate articles to be recommended from the image recognition result.

Taking the video form as an example, carrying out image recognition on each frame of image in the live broadcast forecast information to obtain an image recognition result corresponding to each frame of image; and extracting introduction information of a plurality of candidate articles to be recommended from the image recognition result.

Taking the presentation form of the live broadcast forecast information as an example, carrying out voice recognition on the live broadcast forecast information to obtain a voice recognition result; and extracting the introduction information of a plurality of candidate articles to be recommended from the voice recognition result.

According to the method and the device for extracting the introduction information of the articles to be recommended from the live broadcast forecast information pre-submitted by the live broadcast room, the process of extracting the introduction information of the articles to be recommended from the live broadcast forecast information pre-submitted by the live broadcast room is convenient and rapid, the loss of computing resources can be reduced, and the speed of presenting the recommended pages subsequently can be improved.

In some embodiments, the scheme corresponding to step S1011 and step S1012 and the scheme corresponding to step S1013 and step S1014 may be implemented separately or may be implemented in conjunction with each other.

As one example, introduction information of an item to be recommended is extracted simultaneously from live content and pre-submitted live preview information of at least one live room in a live platform.

Taking 100 recommendation positions of the articles to be recommended as an example, which are needed in the recommendation page, introduction information of the articles to be recommended, which is 100 in total, can be simultaneously extracted from live broadcast content and live broadcast forecast information respectively. The quantity of the introduction information of the articles to be recommended extracted from the live broadcast content and the quantity of the introduction information of the articles to be recommended extracted from the live broadcast forecast information can meet the preset proportion; the preset ratio may be a default value or a value set by the user.

According to the method and the device for extracting the introduction information of the articles to be recommended, the speed of extracting the introduction information of the articles to be recommended can be improved, and timeliness of the extracted introduction information of the articles to be recommended can be guaranteed.

As another example, introduction information of an item to be recommended is extracted from live broadcast forecast information pre-submitted by at least one live broadcast room in a live broadcast platform; and when the quantity of the introduction information of the articles to be recommended is extracted from the live broadcast forecast information and is lower than a quantity threshold value, extracting the introduction information of the articles to be recommended from live broadcast contents of the live broadcast room meeting the interested condition.

Here, the condition of interest includes at least one of: a live room having an interactive relationship (e.g., collection, browsing, purchasing, praying or commenting, etc.) with the live account; live broadcast rooms with interactive relation with live account numbers belong to the same type of live broadcast room; and the live broadcasting room recommended by the account with social relation with the live broadcasting account. The quantity threshold may be a default value or a value set by the user, for example, the quantity threshold may be the number of recommended bits of the item to be recommended required in the recommended page.

Taking 100 recommendation bits of the articles to be recommended as an example, which are needed in the recommendation page, extracting the introduction information of the articles to be recommended from the live broadcasting forecast information preferentially, and continuously extracting the introduction information of 20 articles to be recommended from the live broadcasting contents of the live broadcasting room meeting the interested condition when the quantity of the extracted introduction information of the articles to be recommended is less than 100 (for example, only 80) so that the quantity of the finally obtained introduction information of the articles to be recommended is 100.

Or extracting introduction information of 100 articles to be recommended from the live broadcast forecast information; since the introduction information of the item to be recommended, in which the difference between the predetermined time and the current time (i.e., the time at which the introduction information of the item to be recommended is presented in the recommendation page) exceeds the time threshold, is stale, the introduction information of the item to be recommended, in which the difference between the predetermined time and the current time exceeds the time threshold (for example, the introduction information of 20 items to be recommended, is filtered out), and at this time, the number of the filtered introduction information of the item to be recommended is less than 100, and the extraction of the introduction information of 20 items to be recommended from the live content of the live broadcast room satisfying the condition of interest is continued so that the number of the finally acquired introduction information of the item to be recommended is 100.

In the embodiment of the application, since the introduction information of the article to be recommended is extracted from the live broadcast advance notice information submitted in advance, compared with the introduction information of the article to be recommended from the live broadcast content, the introduction information of the article to be recommended is extracted preferentially from the live broadcast advance notice information, and then the introduction information of the article to be recommended is extracted from the live broadcast content, so that the timeliness of the extracted introduction information of the article to be recommended can be ensured while the speed of extracting the introduction information of the article to be recommended is improved, thereby saving the calculation resources in the recommendation process and improving the speed of presenting the recommendation page.

As yet another example, introduction information of an item to be recommended is extracted in live content of a live room satisfying a condition of interest; when the quantity of the introduction information of the to-be-recommended item extracted from the live broadcast content is lower than the quantity threshold, the introduction information of the to-be-recommended item is extracted from live broadcast advance notice information pre-submitted by at least one live broadcast room in the live broadcast platform (for example, a live broadcast room in the live broadcast platform, from which the introduction information of the to-be-recommended item is not extracted).

Here, the quantity threshold may be a default value or a value set by the user, for example, the quantity threshold may be the number of recommended bits of the item to be recommended required in the recommended page.

Taking 100 recommendation bits of the articles to be recommended as an example, which are needed in the recommendation page, preferentially extracting the introduction information of all the articles to be recommended from the live broadcast content of the live broadcast room meeting the interested condition, and continuously extracting the introduction information of 20 articles to be recommended from the live broadcast forecast information submitted by the live broadcast room from which the introduction information of the articles to be recommended is not extracted when the quantity of the extracted introduction information of the articles to be recommended is less than 100 (for example, only 80) so that the quantity of the finally acquired introduction information of the articles to be recommended is 100.

In the embodiment of the application, compared with the introduction information of the articles to be recommended extracted from the live broadcast content of the live broadcast room meeting the interested condition, the introduction information of the articles to be recommended extracted from the live broadcast forecast information submitted in advance has higher timeliness and accords with the image of the user, so that the introduction information of the articles to be recommended is preferentially extracted from the live broadcast content of the live broadcast room meeting the interested condition, and then the introduction information of the articles to be recommended is extracted from the live broadcast forecast information submitted by the live broadcast room not extracting the introduction information of the articles to be recommended, so that the comprehensiveness and timeliness of the extracted articles to be recommended can be ensured, the recommendation accuracy can be improved, and the drainage probability is improved.

In some embodiments, the second filtering may be further performed on the introduction information of the plurality of articles to be recommended extracted in the above two embodiments, so after extracting the introduction information of the plurality of articles to be recommended, the method may further include:

as one example, determining a recommendation index for each item to be recommended based on the item characteristics in the introduction information; and sorting the plurality of articles to be recommended in a descending order of the recommendation indexes, and filtering out introduction information of the articles to be recommended in the later part of the descending order. In this way, the user can be recommended with a higher preference.

Here, the item characteristics may be used to describe information of the item value, and the recommendation index may be used to measure the recommendation value of the item.

Taking the commodity as an example, the characteristic of the commodity can be information such as price, good score, discount and the like, and the recommendation index can be information such as cost performance and the like. A negative correlation between the recommendation index and the price, a positive correlation (e.g., in a positive proportional relationship) between the recommendation index and the good score, and a positive correlation between the recommendation index and the discount.

As another example, among the introduction information of the plurality of items to be recommended, the introduction information of the item that does not satisfy the condition of interest is filtered out. In this way, items conforming to the interest list can be recommended to the user.

Here, the condition of interest includes at least one of: items having an interactive relationship (e.g., collection, browsing, purchasing, praying or commenting, etc.) with the live account; the articles with interactive relation with the live account belong to the same type of articles; the main broadcast focused by the live account recommends articles in the live broadcast room; and (3) recommending the articles by the account with social relation with the live account.

As yet another example, a predetermined time for each item to be recommended is obtained; among the plurality of articles to be recommended, the introduction information of the articles to be recommended, of which the predetermined time is outside the target time range, is filtered out.

Here, the predetermined time may be an open time of a live presentation of the item, or an open time of an associated merchant beginning to sell the item, i.e., a time at which the item is available (e.g., purchased or reserved). The type of target time range includes at least one of: the expected viewing time range of the live account; idle viewing time range of live account; the time range set by the system is recommended. In this way, the user can be recommended to the item that meets the user's needs for a predetermined time.

For example, the desired viewing time range of the live account may be a time range set by the live account, or may be a time range determined by artificial intelligence technology according to historical viewing data of the live account. Similarly, the idle viewing time range of the live account may also be a time range determined by artificial intelligence technology according to historical viewing data of the live account.

In the following, a specific implementation of the time range determined by artificial intelligence techniques from historical viewing data of a live account is described.

For example, historical viewing data of a live account is obtained; invoking the neural network model to perform the following processing on the historical viewing data: extracting feature vectors of historical viewing data; the extracted feature vectors are mapped to probabilities corresponding to time ranges of a plurality of candidates, respectively, and the time range of the candidate corresponding to the maximum probability is determined as the desired viewing time range. Wherein the sample historical viewing data for training the neural network model comprises: historical viewing time ranges.

In this way, the introduction information of the articles to be recommended in the range of the expected watching time of the user can be accurately acquired through machine learning, so that the recommending efficiency and accuracy can be improved.

In step S102, a recommendation page of the live platform is presented in the human-computer interaction interface.

Here, the recommendation page includes: introduction information of at least one item to be recommended, and a live room entry corresponding to each item.

In some embodiments, presenting the recommended page of the live platform in the human-computer interaction interface may be achieved by: presenting a live broadcasting room display page of a live broadcasting platform in a man-machine interaction interface; and responding to the recommendation operation received in the live broadcasting room display page, and presenting the recommendation page.

Here, the live broadcast room presentation page includes at least one live broadcast room. The recommendation page includes: and displaying the introduction information of the articles to be recommended in at least one live broadcast room displayed in the pages in the live broadcast room.

As an example, the recommended operation may be various forms of operation that are preset by the operating system and that do not conflict with registered operations; or may be a user-defined operation of various forms that is non-conflicting with the registered operation. The recommended action includes at least one of: click operations (e.g., trigger operations for a recommendation button that is presented in a live room presentation page); a sliding operation according to a specific trajectory or direction; voice operation; motion sensing operation (e.g., up-and-down shaking operation or curve motion operation). Thus, the operation experience of the user can be improved.

As an example, the recommended page and the live room presentation page may be displayed simultaneously, e.g., the recommended page and the live room presentation page are displayed on a split screen; and displaying the recommended page above the live broadcasting room display page in a floating layer mode, wherein the recommended page has transparency and can not completely shield the live broadcasting room display page. Of course, the recommendation page and the live room presentation page may not be displayed simultaneously, e.g., when the client switches from the live room presentation page to presenting the recommendation page in response to a recommendation operation.

For example, in fig. 9A and 9B, when the user clicks the "view more" button (i.e., the recommendation button described above) in the live seconds and charts entry 901 in fig. 9A, a live seconds and charts detail page (i.e., the recommendation page described above) corresponding to fig. 9B is presented.

According to the method and the device for recommending the articles, the introduction information of the articles to be recommended in all living broadcast rooms in the living broadcast platform is presented in the recommendation page, the recommendation range can be improved, and a user can intuitively acquire all the articles to be recommended in the current living broadcast platform, so that the recommendation efficiency and accuracy are improved.

In other embodiments, presenting the recommended page of the live platform in the human-computer interaction interface may be implemented by: presenting live broadcast content of a live broadcast room in a man-machine interaction interface; in response to a recommendation operation received in the live room, a recommendation page is presented.

Here, the recommendation page includes: introduction information of items to be recommended in the living room.

As an example, live content of a live room is presented in a human-computer interaction interface, wherein the live room comprises a recommendation button; and responding to the trigger operation for the recommendation button received in the live broadcasting room, and presenting a recommendation page.

According to the method and the device for recommending the articles, the introduction information of the articles to be recommended in the live broadcasting room which the user enters is presented in the recommendation page, the articles recommended in the live broadcasting room which the user is interested in can be presented in a targeted mode, recommending efficiency and accuracy can be improved, and accordingly drainage possibility can be improved.

In some embodiments, when the recommendation page includes introduction information of a plurality of items to be recommended, the ranking manner of the introduction information of the plurality of items to be recommended includes at least one of:

mode one: and sorting the introduction information of the plurality of articles to be recommended according to the sequence of the scheduled time of the articles to be recommended. In this way, the first-order item can be recommended to the user with priority.

For example, in fig. 9B, the merchandise is presented in a time axis 908 in a sequential order from first to last according to the predetermined time of the merchandise.

Mode two: and sorting the introduction information of the plurality of articles to be recommended in descending order according to the interaction heat of the articles to be recommended. In this way, the items with higher interaction heat can be recommended to the user preferentially.

For example, the interactive heat may be measured by information such as sales, click rate, or collection rate of the item to be recommended. The interactive heat and sales are positively correlated, the interactive heat and click rate are positively correlated, and the interactive heat and collection rate are positively correlated.

Mode three: and sorting the introduction information of the plurality of articles to be recommended in a descending order according to the preference degree of the articles to be recommended according to the live account number. In this way, items conforming to the interest list can be recommended preferentially to the user.

For example, the neural network model is invoked to perform the following processing on the introduction information of each item to be recommended: extracting feature vectors of introduction information of articles to be recommended; and mapping the extracted feature vectors into probabilities corresponding to the preference degrees of the plurality of candidates respectively, and determining the preference degree of the candidate corresponding to the maximum probability as the preference degree of the item to be recommended according with the live account.

Here, the neural network model is obtained by training the introduction information of the sample article and the preference degree of the label of the introduction information of the article as the sample, and the neural network model can learn the relation (embodied as the weight parameter of the neural network model) between the characteristics of the introduction information and different preference degrees, so that the probability corresponding to different preference degrees can be predicted for the introduction information of the article to be recommended.

Therefore, the introduction information of the articles to be recommended, which meets the requirements of the user, can be accurately determined through machine learning, and the recommendation efficiency and accuracy can be improved.

Mode four: and sorting the introduction information of the plurality of articles to be recommended in a descending order according to the recommendation indexes of the articles to be recommended. In this way, the user can be preferentially recommended with the item having the higher preference.

For example, the recommendation index may be information such as cost performance. Negative correlation between the recommendation index and the price, positive correlation between the recommendation index and the praise rate, and positive correlation between the recommendation index and the discount.

In some embodiments, the recommendation page further includes a purchase portal corresponding to each item; as such, after step S102, it may further include: in response to a trigger operation for a purchase portal of the target item, a purchase page for purchasing the target item is displayed, thereby facilitating the user's purchase of the target item.

Here, the target item is an item selected among the at least one item to be recommended.

According to the method and the device for purchasing the goods, the corresponding purchasing entrance is presented while the introduction information of the goods is presented, and therefore operation steps of a user can be reduced, operation experience of the user can be improved, and loss of terminal resources can be reduced.

In some embodiments, referring to fig. 7, fig. 7 is a schematic flow diagram of an information recommendation method in live broadcast provided in the embodiment of the present application, and based on fig. 4, step S103 may be further included after step S102.

In step S103, in response to a trigger operation for the live room entrance of the target item, a live room for introducing the target item is displayed.

Here, the target item is an item selected among the at least one item to be recommended. The live room for introducing the target item is a live room where the pointer has a recommendation for the target item about to be, past or present.

In some embodiments, when a plurality of live rooms for introducing the target object exist, after receiving a trigger operation of a live room entrance for the target object, presenting a live room list, wherein the live room list presents the plurality of live rooms for introducing the target object; and displaying the live broadcasting room selected by the live broadcasting room selection operation in response to the live broadcasting room selection operation received in the live broadcasting room list.

In some embodiments, the live room for introducing the target item is displayed when the difference between the predetermined time of the target item and the trigger operation time does not exceed a time difference threshold.

Here, the time difference threshold may be a default value or a value set by the user.

For example, in fig. 9B, if the predetermined time interval of the target item is less than half an hour apart from the current time (i.e., the time when the introduction information of the item to be recommended is presented in the recommendation page), indicating that the live broadcast is about to start, the user may directly enter the live broadcast room to watch by triggering the on-hook button 902.

In other embodiments, when the difference between the predetermined time of the target item and the trigger operation time exceeds a time difference threshold, the user is automatically or prompted to generate a reminder task and the reminder task is executed before the predetermined time to present reminder information.

For example, in fig. 9B, if the predetermined time interval of the target item is greater than half an hour apart from the current time (i.e., the time when the introduction information of the item to be recommended is presented in the recommendation page), the user may set the preemptive reminder by triggering the reminder button 903.

As one example, when the human-machine interface is presented with a live room for introducing the target item, a purchase portal for the target item is presented in the live room in a popup window or floating layer.

For example, in fig. 9F, when the user has logged into the live broadcast room and detected that the commodity of interest to the user is about to be robbed for the first 30s, the server sends a reminder message to the client, so that the client presents the reminder 907 in the form of a pop-up window, notifying the user that the user can enter the purchase page of the commodity to rob.

As another example, when the human-machine interaction interface presents a live room and the live room is not for introducing the target item, the live client may present the live room portal of the target item in a popup or floating layer manner in the live room.

For example, when a user is watching a live broadcast of a recommended item a and an item B of a reminder task is set to be about to reach a robbery time, a live broadcast room entrance of the item B is presented in a live broadcast room in a popup window form so that the user can quickly jump to the live broadcast room of the recommended item B. Thus, the user can be prevented from missing the available time of the articles to be recommended, and the purchasing experience of the user is improved.

As yet another example, when the human-machine interface is not presented with a live room, the live room portal and/or purchase portal is presented in the form of a notification message.

For example, fig. 9D is 5 minutes before the product is robbed, and the system notifies the user, via the issued applet reminder 905, that the user can enter the live room to watch live and pay attention to the product of interest.

According to the method and the device for reminding the user, the user is supported to set reminding before the preset time of the article to be recommended, the user can be prevented from missing the available time of the article to be recommended, and therefore purchasing experience of the user is improved.

Next, an information recommendation method in live broadcast provided in the embodiment of the present application will be described by taking a cooperation of the terminal 200 and the server 400 in fig. 2 as an example. Referring to fig. 8, fig. 8 is a flowchart of a method for recommending information in live broadcast according to an embodiment of the present application, and the steps shown in fig. 8 will be described.

In step S801, the server extracts introduction information of an item to be recommended from live content of at least one live room in the live platform and live preview information submitted in advance.

In step S802, the server transmits introduction information of an item to be recommended to the client.

In step S803, the client presents introduction information of the item to be recommended and a live room entry corresponding to each item in the recommendation page of the live platform.

It should be noted that, the specific implementation manner of the steps S801 to S803 is similar to the embodiments contained in the steps S101 to S103, and will not be described herein.

In the embodiment of the invention, the server has strong computing capability and high operation speed relative to the terminal, and the process of extracting the introduction information of the articles to be recommended is completed through the server, so that the speed of presenting the introduction information of the articles to be recommended in the recommendation page by the terminal can be improved, and the computing resource of the terminal can be reduced.

The method for recommending information in live broadcast provided by the embodiment of the application is described below by taking the commodity as an example.

The method and the device are mainly applied to live broadcast scenes of electronic commerce, all recently-opened and forecast commodities of a platform (namely the live broadcast platform) can be pulled in real time by the background, all the live broadcast second-to-be-killed commodities of the live broadcast in all the live broadcast rooms of the platform are detected in real time based on voice recognition, the price and preferential strength of the broadcast forecast commodities and the live broadcast second-to-be-killed commodities of the live broadcast in the live broadcast rooms are compared with similar commodity prices and preferential strengths, commodities with higher intelligent screening performance price are generated, and a carefully chosen live broadcast second-to-killing list (namely the recommended page) is generated. Thus, when a user enters the APP home page, the user can check the corresponding live broadcast and purchase through the list of commodity dimension. According to the method and the device for achieving the direct broadcast commodity, the audience can be helped to rapidly reach the direct broadcast commodity with high cost performance, the user experience of the audience is improved, and the popularity of middle and small merchants can be improved through a decentralization distribution mode.

It should be noted that, the background and the server in this example are different implementation forms of the server 200 in fig. 2, and the background corresponds to a node for data processing, and supports interaction with operation and maintenance personnel; the servers correspond to nodes for data storage and distribution.

Referring to fig. 9A, 9B, 9C, 9D, 9E, and 9F, fig. 9A, 9B, 9C, 9D, 9E, and 9F are application scene diagrams of the information recommendation method in live broadcasting provided in the embodiments of the present application.

Fig. 9A is a home page of an e-commerce live, where there is a live seconds listing entry 901 at the top of the home page, where the live seconds listing entry 901 is aggregated with the goods and prices as dimensions. FIG. 9B is a live seconds list details page (i.e., the recommended page described above) entered when the user clicks the "view more" button in the live seconds list entry 901, where the details page presents items in a timeline, and the details page highlights higher cost-performance seconds of the items, if the time of opening (i.e., the predetermined time described above) is less than half an hour from the current time (i.e., the time of presenting items in the live seconds list), the user can directly enter the live broadcasting room to view by triggering the open robber button 902; if the time of opening is greater than half an hour from the current time (i.e., the time of presenting the merchandise in the live second leadership) the user can set an opening reminder by triggering the reminder button 903.

Fig. 9C is a pop-up window page in which the user sets the tamper notification, informing the user in the pop-up window 904 that the tamper notification has been set. 5 minutes before the main broadcasting introduces the commodity, the background reminds the user to enter the live broadcasting room for watching through a live broadcasting client side in a small program and short message mode. Fig. 9D shows that 5 minutes before the product is robbed, the background notifies the user that the user can enter the live broadcasting room to watch the live broadcasting and pay attention to the product of interest by sending the applet reminding information 905 to the live broadcasting client, wherein the sending of the short message is in a common short message mode.

Fig. 9E is a live broadcast room with an E-commerce, with a commodity floating layer in the lower left corner of the live broadcast room being currently being taught by the host and a purchase portal 906 for the live broadcast room commodity. In fig. 9F, when the user has entered the live broadcast room and detects the first 30s that the commodity focused by the user is about to be robbed, the background will send a reminding message to the live broadcast client, so that the live broadcast client presents the reminding message 907 in the form of a popup window, and notifies the user that the user can enter the commodity detail page (i.e. the purchasing page of the commodity) to rob.

Next, a specific implementation manner of the embodiment of the present application will be described with reference to fig. 10, and fig. 10 is a schematic flow chart of an information recommendation method in live broadcast provided in the embodiment of the present application.

In step S110, the user enters the live home page to view a live seconds killing bar.

In some embodiments, the information in the live second killing list is generated by the background based on the price of the commodity information in all live previews of the platform and the price of the commodity information in the live content for which the upcoming second killing is detected.

In step S120, the background pulls information of all live preview products of the platform.

In some embodiments, the background will pull the names and prices of the merchandise hung in all live trailers (i.e., live trailer information described above) on the platform every second.

In step S130, the client speech identifies and monitors all the live broadcasting room anchor' S mouth-casts in real time, and generates the commodity and price to be killed in seconds.

In some embodiments, after the anchor is opened, the client monitors the mouth-casting of all the living rooms of the platform in real time, and generates the commodity and the price to be killed in seconds based on the voice information of the mouth-casting.

Next, a specific implementation manner of the voice recognition platform for broadcasting the mouth of the living broadcasting room and generating the goods and prices to be killed in seconds will be described with reference to fig. 11 and 12, fig. 11 is a schematic flow chart of voice recognition provided in the embodiment of the present application, and fig. 12 is a schematic diagram of voice recognition provided in the embodiment of the present application.

1) Preprocessing and feature extraction of anchor speech

Before speech recognition begins, the speech needs to be analyzed by silence removal operation (VAD, voice Activi ty Detection), the anchor speech is framed and cut into small segments, each of which is called a frame, and the frames are overlapped, as shown in fig. 12, and the silence of the head and tail ends is removed to reduce the interference caused to the subsequent steps. In fig. 12, each frame has a length of 25 milliseconds (ms), and there is 25-10=15 ms overlap between each two frames. After framing, the speech becomes a plurality of small segments, and then, each frame of waveform is changed into a multidimensional vector according to the physiological characteristics of human ears, and acoustic feature extraction is performed to obtain feature vectors (i.e., the above-mentioned audio vectors) of the input speech.

2) Similarity measurement and post-processing

Firstly, detecting the dynamic state of a commodity floating layer in the live broadcast field in real time by a background, and automatically identifying commodity texts in the floating layer when the commodity floating layer pops up; then, a dynamic time warping algorithm (DTW, dynamicTi me Waplng) is applied for speech recognition. Specifically, the feature vector after the preprocessing and feature extraction is compared with each template in the reference voice template library (i.e., the reference template described above) in sequence to obtain the similarity between them, and the highest similarity is output as the recognition result.

Here, the reference voice template library is obtained by matching the commodity document in the floating layer with the basic template in the basic voice template library. The base templates included in the base speech template library are feature vectors corresponding to the speech of the words in the base vocabulary.

When the voice of the anchor is detected, the audio data are collected and sent to the background, after the background receives the audio data, the audio is decompressed and transcoded, and is subjected to similarity matching with a reference voice template library, when the content of the anchor is detected to contain time, second killing, commodity or price related vocabulary for 2 times within a certain time threshold (for example, 3-5 min), the audio data after the characteristics are extracted are subjected to text output through an acoustic model, a dictionary and a language model.

It should be noted that, the step S120 and the step S130 may be performed in parallel, or may be performed in no sequence.

In step S140, the background uploads the collected live broadcast forecast and the commodity information in the live broadcast to the server, and the linked big data screening is lower than the price of the whole network commodity.

In some embodiments, the client requests CDN data from the background, pulls big data from the cloud repository, links with an artificial intelligence (AI, artificial Intelligence) technology, queries prices of similar commodities on the network, matches and checks the prices of similar commodities with prices of the commodities on the live platform, and returns information of the commodity and the anchor's time to the background when detecting that the price of the commodity is the lowest price on the whole network. The client requests the data from the background every second to ensure the real-time update of the data of the list.

In step S150, the client generates a selected live second killing list, and determines a start time.

In some embodiments, the client generates a selected live broadcast second killing list based on the data returned by the background, and judges the time of live broadcast on the second killing list, and when the time of commodity robbing is less than half an hour, the user clicks the card and can directly enter the live broadcast room to watch live broadcast. When the time for opening the commodity is more than half an hour, the user can choose to enter the live broadcasting room for watching, and can also choose to open the robbing prompt. And 5 minutes before the commodity starts to rob (the reminding time can be set by a user and can be any time length, and the limitation is not made here), and the background sends the commodity starting and rob reminding to the client of the corresponding user through a small program and a short message. After the user enters the live broadcasting room, 30s before the second killed commodity is robbed, the client side can pop up popup window information of the commodity to remind the user of entering the detail page to prepare for robbed.

An exemplary structure of the in-live information recommendation device 555 implemented as a software module provided in the embodiments of the present application is described below with reference to fig. 3, and in some embodiments, as shown in fig. 3, the software module stored in the in-live information recommendation device 555 of the memory 550 may include:

The acquisition module 5551 is configured to acquire introduction information of an item to be recommended in a live broadcasting room of the live broadcasting platform;

the live broadcast presenting module 5552 is configured to present a recommendation page of the live broadcast platform in the human-computer interaction interface;

the recommendation page comprises the following steps: introduction information of at least one item to be recommended, and a live room entry corresponding to each item.

In the above scheme, the obtaining module 5551 is further configured to perform real-time voice recognition on live content in at least one live broadcast room in the live broadcast platform, so as to obtain a voice recognition result; and extracting introduction information of a plurality of articles to be recommended from the voice recognition result.

In the above scheme, the obtaining module 5551 is further configured to extract audio data from live broadcast content, and perform frame processing on the audio data to obtain a plurality of audio sub-data; extracting acoustic features of each audio sub-data to obtain a plurality of audio vectors; the following processing is performed for each audio vector: matching the audio vector with each reference template in the reference voice template library to determine the similarity between the audio vector and each reference template, and determining the text information corresponding to the reference template with the highest similarity as the text information corresponding to the audio vector; and combining a plurality of text information corresponding to the plurality of audio sub-data one by one to obtain a voice recognition result.

In the above scheme, the obtaining module 5551 is further configured to obtain text information in a floating layer included in the live broadcast room; and in the basic voice template library, determining a plurality of basic templates matched with the text information in the floating layer, and combining the determined basic templates into a reference voice template library.

In the above scheme, the obtaining module 5551 is further configured to obtain live broadcast forecast information pre-submitted by at least one live broadcast room in the live broadcast platform; and extracting introduction information of a plurality of articles to be recommended from the live broadcast forecast information.

In the above solution, the obtaining module 5551 is further configured to determine a recommendation index of each item to be recommended according to the item feature in the introduction information; and sorting the plurality of articles to be recommended in a descending order of the recommendation indexes, and filtering out introduction information of the articles to be recommended in the later part of the descending order.

In the above solution, the obtaining module 5551 is further configured to filter out, from the introduction information of the plurality of articles to be recommended, the introduction information of the articles that do not satisfy the condition of interest; wherein the condition of interest comprises at least one of: items having an interactive relationship with the live account; the articles with interactive relation with the live account belong to the same type of articles; the main broadcast focused by the live account recommends articles in the live broadcast room; and (3) recommending the articles by the account with social relation with the live account.

In the above-mentioned scheme, the acquiring module 5551 is further configured to acquire a predetermined time of each item to be recommended; filtering out introduction information of the articles to be recommended, wherein the preset time of the introduction information is out of the target time range, from the plurality of articles to be recommended; wherein the type of the target time range includes at least one of: the expected viewing time range of the live account; idle viewing time range of live account numbers.

In the above solution, the live presentation module 5552 is further configured to display a live room for introducing the target item in response to a trigger operation of the live room portal for the target item; wherein the target item is a selected item among the at least one item to be recommended.

In the above solution, the live presentation module 5552 is further configured to display a live room for introducing the target item when a difference between the predetermined time of the target item and the trigger operation time does not exceed a time difference threshold; when the difference between the preset time of the target object and the trigger operation time exceeds a time difference threshold, generating a reminding task, and executing the reminding task before the preset time to present reminding information.

In the above solution, the live broadcast presenting module 5552 is further configured to present, when the human-computer interaction interface presents a live broadcast room for introducing the target object, a purchase entrance of the target object in the live broadcast room in a popup window or floating layer manner; when the man-machine interaction interface presents a live broadcasting room and the live broadcasting room is not used for introducing the target object, presenting a live broadcasting room entrance of the target object in the live broadcasting room in a popup window or floating layer mode; and when the human-computer interaction interface does not present the live broadcasting room, presenting the live broadcasting room entrance and/or the purchase entrance in the form of notification messages.

In the above scheme, the recommendation page further comprises a purchase inlet corresponding to each article; the live presentation module 5552 is further configured to display a purchase page for purchasing the target item in response to a trigger operation of the purchase portal for the target item; wherein the target item is a selected item among the at least one item to be recommended.

In the above scheme, the live broadcast presenting module 5552 is further configured to present a live broadcast room display page of the live broadcast platform in the human-computer interaction interface; responding to the recommendation operation received in the live broadcasting room display page, and presenting a recommendation page; the recommendation page comprises the following steps: and displaying the introduction information of the articles to be recommended in at least one live broadcast room displayed in the pages in the live broadcast room.

In the above scheme, the live broadcast presenting module 5552 is further configured to present live broadcast content in the live broadcast room in the human-computer interaction interface; presenting a recommendation page in response to a recommendation operation received in the live broadcast room; the recommendation page comprises the following steps: introduction information of items to be recommended in the living room.

In the above scheme, when the recommendation page includes introduction information of a plurality of items to be recommended, the sorting manner of the introduction information of the plurality of items to be recommended includes at least one of the following: sorting the introduction information of the plurality of articles to be recommended according to the sequence of the scheduled time of the articles to be recommended; sorting the introduction information of a plurality of articles to be recommended in descending order according to the interaction heat of the articles to be recommended; the method comprises the steps of sorting introduction information of a plurality of articles to be recommended in descending order according to the preference degree of the articles to be recommended according to a live account; and sorting the introduction information of the plurality of articles to be recommended in a descending order according to the recommendation indexes of the articles to be recommended.

Embodiments of the present application 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, so that the computer device executes the information recommendation method in live broadcast according to the embodiment of the application.

The embodiments of the present application provide a computer-readable storage medium storing computer-executable instructions, in which the computer-executable instructions are stored, which when executed by a processor, cause the processor to perform the method for recommending information in live broadcast provided in the embodiments of the present application, for example, the method for recommending information in live broadcast shown in fig. 4, fig. 5, fig. 6, fig. 7, fig. 8, and fig. 10, where the computer includes various computing devices including a smart terminal and a server.

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

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

As an example, computer-executable instructions may, but need not, correspond to files in a file system, may be stored in a portion of a file that holds other programs or data, such as 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).

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

In summary, the embodiment of the application has the following beneficial effects:

(1) And comparing the similarity of each audio vector contained in the audio data extracted from the live broadcast content with a reference template in a reference voice template library in sequence to obtain a voice recognition result, wherein compared with the method of directly searching corresponding text information in a basic voice template library, the voice recognition speed is faster, the accuracy is higher, and therefore the consumption of computing resources can be reduced.

(2) Because the introduction information of the articles to be recommended is extracted from the live broadcast forecast information submitted in advance, compared with the introduction information of the articles to be recommended from the live broadcast content, the introduction information of the articles to be recommended is extracted from the live broadcast forecast information preferentially, so that the calculation resources can be saved, and the speed of presenting the recommended page is improved.

(3) Because the introduction information of the articles to be recommended extracted from the live broadcast content is more in line with the portrait of the user than the introduction information of the articles to be recommended extracted from the live broadcast forecast information submitted in advance, the recommendation accuracy can be improved, and the drainage probability is improved.

(4) The introduction information of the articles to be recommended in all live broadcasting rooms in the live broadcasting platform is presented in the recommendation page, the recommendation range can be improved, and a user can intuitively acquire all the articles to be recommended in the current live broadcasting platform, so that the recommendation efficiency and accuracy are improved.

(5) The introduction information of the articles to be recommended in the live broadcast room which the user enters is presented in the recommendation page, the articles recommended in the live broadcast room which the user is interested in can be presented to the user in a targeted manner, the recommendation efficiency and accuracy can be improved, and therefore the drainage possibility can be improved.

(6) The reminding method and the reminding device support the user to set the reminding before the preset time of the articles to be recommended, and can avoid the user to miss the available time of the articles to be recommended, so that the purchasing experience of the user is improved.

The foregoing is merely exemplary embodiments of the present application and is not intended to limit the scope of the present application. Any modifications, equivalent substitutions, improvements, etc. that are within the spirit and scope of the present application are intended to be included within the scope of the present application.

Claims

1. An information recommendation method in live broadcasting is characterized by comprising the following steps:

responding to a triggering operation of a live broadcasting room entrance aiming at a target object, displaying a live broadcasting room for introducing the target object when the difference value between the preset time of the target object and the triggering operation time does not exceed a time difference threshold value, generating a reminding task when the difference value between the preset time of the target object and the triggering operation time exceeds a time difference threshold value, and executing the reminding task before the preset time to present reminding information;

Wherein, the recommendation page includes: the target item is selected from the at least one item to be recommended, and the entry of the living broadcasting room corresponding to each item, wherein the preset time is the opening time of living broadcasting introducing the item to be recommended or the opening time of starting selling the item to be recommended.

2. The method according to claim 1, wherein the obtaining introduction information of the item to be recommended in the living room of the living platform includes:

performing real-time voice recognition on live contents of at least one live broadcasting room in the live broadcasting platform to obtain a voice recognition result;

and extracting introduction information of a plurality of articles to be recommended from the voice recognition result.

3. The method according to claim 2, wherein performing real-time voice recognition on live content of at least one live room in the live platform to obtain a voice recognition result comprises:

extracting audio data from the live broadcast content, and framing the audio data to obtain a plurality of audio sub-data;

extracting acoustic features of each audio sub-data to obtain a plurality of audio vectors;

The following is performed for each of the audio vectors:

matching the audio vector with each reference template in a reference voice template library to determine the similarity between the audio vector and each reference template, and determining the text information corresponding to the reference template with the highest similarity as the text information corresponding to the audio vector;

and combining a plurality of text information corresponding to the audio sub-data one by one to obtain the voice recognition result.

4. A method according to claim 3, wherein prior to said matching the audio vector with each reference template in a library of reference speech templates, the method further comprises:

acquiring character information in a floating layer included in the live broadcasting room;

and in the basic voice template library, determining a plurality of basic templates matched with the text information in the floating layer, and combining the determined basic templates into the reference voice template library.

5. The method according to claim 1, wherein the obtaining introduction information of the item to be recommended in the living room of the living platform includes:

acquiring live broadcast forecast information pre-submitted by at least one live broadcast room in the live broadcast platform;

And extracting introduction information of a plurality of articles to be recommended from the live broadcast forecast information.

6. The method according to claim 2 or 5, wherein after said extracting introduction information of a plurality of items to be recommended, the method further comprises:

determining a recommendation index of each item to be recommended according to the item characteristics in the introduction information;

and sorting the plurality of articles to be recommended in a descending order of recommendation indexes, and filtering out introduction information of the articles to be recommended in the later part of the descending order.

7. The method according to claim 2 or 5, wherein after said extracting introduction information of a plurality of items to be recommended, the method further comprises:

filtering out the introduction information of the articles which do not meet the interested conditions from the introduction information of the articles to be recommended;

wherein the condition of interest comprises at least one of:

items having an interactive relationship with the live account; the main broadcast focused by the live account recommends articles in the live broadcast room; and (3) recommending the articles by the account with social relation with the live account.

8. The method according to claim 2 or 5, wherein after said extracting introduction information of a plurality of items to be recommended, the method further comprises:

Acquiring the preset time of each article to be recommended;

filtering out introduction information of the articles to be recommended, wherein the preset time of the introduction information is out of a target time range, from a plurality of articles to be recommended;

wherein the type of the target time range includes at least one of:

the expected viewing time range of the live account; idle viewing time range of live account numbers.

9. The method of claim 1, wherein presenting the reminder information comprises:

when the man-machine interaction interface presents a live broadcasting room for introducing the target object, presenting a purchase entrance of the target object in the live broadcasting room in a popup window or floating layer mode;

when the man-machine interaction interface presents a live broadcasting room and the live broadcasting room is not used for introducing the target object, presenting a live broadcasting room entrance of the target object in the live broadcasting room in a popup window or floating layer mode;

and when the man-machine interaction interface does not present the live broadcasting room, presenting the live broadcasting room entrance and/or the purchase entrance in a notification message mode.

10. The method of claim 1, wherein when the recommendation page includes introduction information of a plurality of the items to be recommended, the manner of ordering the introduction information of the plurality of the items to be recommended includes at least one of:

Sorting the introduction information of the plurality of articles to be recommended according to the sequence of the preset time of the articles to be recommended;

sorting the introduction information of the plurality of articles to be recommended in a descending order according to the interaction heat of the articles to be recommended;

sorting the introduction information of the plurality of articles to be recommended in a descending order according to the preference degree of the articles to be recommended according with the live account;

and sorting the introduction information of the plurality of articles to be recommended in a descending order according to the recommendation indexes of the articles to be recommended.

11. An information recommendation device in live broadcast, characterized by comprising:

the live broadcast presenting module is used for presenting a recommended page of the live broadcast platform in the man-machine interaction interface; responding to a triggering operation of a live broadcasting room entrance aiming at a target object, displaying a live broadcasting room for introducing the target object when the difference value between the preset time of the target object and the triggering operation time does not exceed a time difference threshold value, generating a reminding task when the difference value between the preset time of the target object and the triggering operation time exceeds a time difference threshold value, and executing the reminding task before the preset time to present reminding information;

12. An electronic device, comprising:

a memory for storing computer executable instructions;

a processor for implementing the method for recommending information in live broadcast according to any of claims 1 to 10 when executing computer executable instructions stored in said memory.

13. A computer readable storage medium, characterized in that computer executable instructions are stored, which when executed are adapted to implement the method of information recommendation in live broadcast according to any of claims 1 to 10.