Method and system for automatically complementing on-demand metadata information

Abstract

The invention is suitable for the technical field of on-demand, and provides an on-demand metadata information automatic complementing method and system. According to the method, data collection and preprocessing are carried out on existing metadata information of stock media assets, a stock media asset completion model is constructed, and information missing completion is carried out; constructing a hotspot media asset completion model, and performing data integration and AI auxiliary completion with the missing completion metadata to obtain integrated completion metadata; establishing a feature library of important contents of holidays and festivals and society, carrying out data feature association, and carrying out specific media asset recommendation; and constructing a user behavior portrait and a user personalized recommendation model, and generating a personalized recommendation list for the user. According to the technical scheme of the invention, by complementing the detailed metadata of the stock media assets, a user can obtain a more accurate result when searching and screening contents and obtain hot contents in time, so that the user can obtain hot contents and related information thereof in time on an on-demand platform, and more accurate personalized and situational recommendation can be realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of video on demand, and in particular relates to a method and system for automatically completing video on demand metadata information. Background Art

[0002] The existing on-demand technology has the following technical defects:

[0003] (1) Incomplete metadata of existing media assets: In the on-demand system, there are a large number of existing media assets. However, the metadata of these media assets often contains missing information. For example, the metadata of some films may only contain basic names and durations, but lack detailed information such as directors, actors, and genre breakdowns. This makes it impossible for users to obtain accurate references when searching and filtering content, affecting the user experience;

[0004] (2) Hot media data are not updated in a timely manner: With the continuous emergence of hot content on the Internet, there is a lag in the replenishment of hot media data in the on-demand system. For example, when a popular TV series arouses widespread discussion on the Internet, the relevant metadata in the on-demand system may not be updated in time with the TV series' popularity ranking, hot topic associations and other information, resulting in the inability to effectively recommend hot content to users;

[0005] (3) Lack of personalized and differentiated operations: Most existing on-demand recommendation systems are based on general algorithms and do not fully consider the personalized needs of users. At the same time, during different holidays or special periods, they lack the ability to make targeted recommendations based on important social content and cannot achieve differentiated automatic operations. For example, during the Spring Festival, there are no recommendations for special content with Spring Festival themes, such as Spring Festival Gala-related programs, documentaries on traditional Spring Festival customs, etc.;

[0006] (4) Insufficient data comparison and completion potential: Although AI technology has made significant progress in the field of data processing, it has not yet been fully applied in the completion of on-demand metadata;

[0007] (5) Personalized recommendations are not accurate enough: In terms of personalized recommendations, existing on-demand systems often make recommendations based on simple user viewing history, and do not deeply explore the ability of AI technology to analyze multi-dimensional data such as user behavior and preferences to achieve more accurate personalized recommendations. Summary of the invention

[0008] The purpose of the embodiments of the present invention is to provide a method and system for automatically completing on-demand metadata information, aiming to solve the technical problems existing in the prior art mentioned in the background technology.

[0009] The embodiment of the present invention is implemented as follows:

[0010] A method for automatically completing on-demand metadata information, the method specifically comprising the following steps:

[0011] Collect and preprocess the existing metadata information of stock media assets, build a stock media asset completion model, complete missing information, and obtain missing completion metadata;

[0012] Monitor the hot media information of the entire network in real time, build a hot media completion model, and perform data integration and AI-assisted completion with the missing completion metadata to obtain integrated completion metadata;

[0013] Establish a feature database of holidays and important social content, associate data features, and recommend specific media resources;

[0014] Collect user personalized information, build user behavior portraits and user personalized recommendation models, and generate personalized recommendation lists for users.

[0015] As a further limitation of the technical solution of the embodiment of the present invention, the data collection and preprocessing of the existing metadata information of the stock media assets, the construction of the stock media asset completion model, the completion of missing information, and the acquisition of missing completion metadata specifically include the following steps:

[0016] Collect existing metadata information of stock media assets;

[0017] Cleaning the metadata information to remove erroneous data and data in irregular formats to obtain standard data;

[0018] Build a model to supplement existing media resources;

[0019] The standard data is input into the stock media asset completion model to predict and complete missing information, thereby obtaining missing completion metadata.

[0020] As a further limitation of the technical solution of the embodiment of the present invention, in the construction of the existing media asset completion model, a preset existing media asset metadata sample is used, and hyperparameters are set to perform model training processing on the convolutional neural network or the recurrent neural network until the model converges to generate a stock media asset completion model, and the hyperparameters include a learning rate and a number of iterations.

[0021] As a further limitation of the technical solution of the embodiment of the present invention, the metadata information includes a basic name, duration and release time; the missing completion metadata includes a basic name, duration, release time, director, actor and detailed type.

[0022] As a further limitation of the technical solution of the embodiment of the present invention, the real-time monitoring of the hot media information of the entire network, building a hot media completion model, and performing data integration and AI-assisted completion with the missing completion metadata to obtain the integrated completion metadata specifically include the following steps:

[0023] Through web crawler technology, real-time monitoring of hot media information on the entire network;

[0024] Integrate the hot media asset information with the missing and completed metadata to obtain integrated metadata;

[0025] Construct a hot media resource completion model;

[0026] The integrated metadata is completed with AI assistance through the hot media resource completion model to obtain integrated completed metadata.

[0027] As a further limitation of the technical solution of the embodiment of the present invention, the hot media information includes popularity rankings, hot topics and hot celebrity dynamics.

[0028] As a further limitation of the technical solution of the embodiment of the present invention, the step of integrating the hot media asset information with the missing completion metadata to obtain the integrated metadata specifically includes the following steps:

[0029] According to the hot media asset information, supplement the relevant hot metadata in the missing and completed metadata;

[0030] According to the hot media asset information, a metadata framework is constructed to supplement the hot media assets that are not related to the missing and completed metadata;

[0031] Generates integration metadata.

[0032] As a further limitation of the technical solution of the embodiment of the present invention, the establishment of a feature library of holidays and important social content, data feature association, and specific media asset recommendation specifically include the following steps:

[0033] Establishing a feature database of holidays and important social contents, wherein the feature database includes family reunion dates, traditional custom dates and national memorial dates;

[0034] Based on the feature library, data feature association is performed on the integrated and completed metadata, and feature association information is recorded;

[0035] Specific media resources are recommended based on the feature association information.

[0036] As a further limitation of the technical solution of the embodiment of the present invention, the collecting of user personalized information, building of user behavior profile and user personalized recommendation model, and generating a personalized recommendation list for the user specifically include the following steps:

[0037] Collect user personalized information, including viewing history, favorites, and search keyword information;

[0038] Building a user behavior profile based on the user personalized information;

[0039] Build a user personalized recommendation model;

[0040] The user behavior profile is input into the user personalized recommendation model to generate a personalized recommendation list for the user.

[0041] A system for automatically completing on-demand metadata information, the system comprising a stock media resource completion module, a hot media resource completion module, a specific media resource recommendation module and a user personalized recommendation module, wherein:

[0042] The stock media asset completion module is used to collect and preprocess the existing metadata information of the stock media assets, build a stock media asset completion model, complete the missing information, and obtain the missing completion metadata;

[0043] A hot media resource completion module is used to monitor the hot media resource information of the entire network in real time, build a hot media resource completion model, and perform data integration and AI-assisted completion with the missing completion metadata to obtain integrated completion metadata;

[0044] The specific media resource recommendation module is used to build a feature library of holidays and important social content, associate data features, and recommend specific media resources;

[0045] The user personalized recommendation module is used to collect user personalized information, build user behavior portraits and user personalized recommendation models, and generate personalized recommendation lists for users.

[0046] Compared with the prior art, the present invention has the following beneficial effects:

[0047] (1) More accurate search and filtering: By completing the detailed metadata of existing media assets, users can obtain more accurate results when searching and filtering content. For example, users can conduct precise searches based on director, actor, detailed type, and other conditions to quickly find content that interests them.

[0048] (2) Ability to obtain hot content in a timely manner: The timely replenishment of hot media data across the entire network enables users to obtain popular content and related information on the on-demand platform in a timely manner, and will not miss out on currently popular TV series, movies or variety shows, etc.

[0049] (3) Achieve more precise personalized and contextual recommendations: Make recommendations based on users’ personalized needs and different holidays and social situations, which improves the match between recommended content and user interests and enhances user satisfaction and stickiness to the platform.

[0050] (4) Capable of differentiated operations: Capable of performing differentiated content recommendations and operations based on different periods and user groups, thus improving the competitiveness of the on-demand platform in the market.

[0051] (5) Improving content utilization: By completing metadata and making accurate recommendations, more existing media resources and hot media resources can be discovered and viewed by users, improving content utilization and reducing the waste of content resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 A flow chart of a method for automatically completing on-demand metadata information provided by an embodiment of the present invention is shown;

[0053] Figure 2 The structure diagram of the on-demand metadata information automatic completion system provided by an embodiment of the present invention is shown. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0055] It is understandable that the existing on-demand technology has the following technical defects: (1) Incomplete metadata of existing media assets: In the on-demand system, there are a large number of existing media assets, however, the metadata of these media assets often lack information. For example, the metadata of some films may only contain basic names and durations, lacking detailed information such as directors, actors, and genre breakdowns. This makes it impossible for users to obtain accurate references when searching and filtering content, affecting user experience; (2) Hot media asset data is not updated in a timely manner: With the continuous emergence of hot content on the Internet, there is a lag in the replenishment of hot media asset data in the on-demand system. For example, when a popular TV series has caused widespread discussion on the Internet, the relevant metadata in the on-demand system may not be updated in time with the popularity ranking of the TV series, hot topic associations and other information, resulting in the inability to effectively recommend hot content to users; (3) Lack of personalized and differentiated operations: Most of the existing on-demand recommendation systems are based on general algorithms and do not fully consider the personalized needs of users. At the same time, during different holidays or special periods, they lack the ability to make targeted recommendations based on important social content and cannot achieve differentiated automatic operations. For example, during the Spring Festival, there are no recommendations for special content with Spring Festival themes, such as Spring Festival Gala-related programs, documentaries on traditional Spring Festival customs, etc.; (4) Insufficient data comparison and completion potential: Although AI technology has made significant progress in the field of data processing, it has not yet been fully applied in the completion of on-demand metadata; (5) Personalized recommendations are not accurate enough: In terms of personalized recommendations, existing on-demand systems often make recommendations based on simple user viewing history, and do not deeply explore the ability of AI technology in analyzing multi-dimensional data such as user behavior and preferences to achieve more accurate personalized recommendations.

[0056] In order to solve the above problems, the embodiment of the present invention discloses a method and system for automatically completing metadata information on demand. The method collects and preprocesses the existing metadata information of stock media assets, builds a stock media asset completion model, completes missing information, and obtains missing completion metadata; monitors the hot media asset information of the entire network in real time, builds a hot media asset completion model, and performs data integration and AI-assisted completion with missing completion metadata to obtain integrated completion metadata; establishes a feature library of holidays and important social content, associates data features, and recommends specific media assets; collects user personalized information, builds user behavior portraits and user personalized recommendation models, and generates personalized recommendation lists for users. By completing the detailed metadata of stock media assets, users can obtain more accurate results when searching and filtering content, and obtain hot content in a timely manner, so that users can obtain popular content and related information on the on-demand platform in a timely manner, and can achieve more accurate personalized and contextual recommendations.

[0057] Specifically, Figure 1 A flow chart of a method for automatically completing on-demand metadata information provided by an embodiment of the present invention is shown.

[0058] In a preferred embodiment of the present invention, a method for automatically completing on-demand metadata information includes the following steps:

[0059] Step S101: collect and preprocess the existing metadata information of the stock media assets, build a stock media asset completion model, complete the missing information, and obtain missing completion metadata.

[0060] In an embodiment of the present invention, metadata information such as the existing basic name, duration, and release time of stock media assets is collected, and then data cleaning is performed on the metadata information to remove erroneous data and data in non-standard formats in the metadata information to obtain standard data. Preset stock media asset metadata samples are used to set hyperparameters such as learning rate and number of iterations, and model training processing is performed on a convolutional neural network or a recurrent neural network until the model converges to generate a stock media asset completion model. The standard data is then input into the stock media asset completion model to predict and complete missing information, thereby obtaining missing completion metadata with basic name, duration, release time, director, actor, and detailed type.

[0061] Step S102: monitor the hot media information of the entire network in real time, build a hot media completion model, and perform data integration and AI-assisted completion with the missing completion metadata to obtain integrated completion metadata.

[0062] In an embodiment of the present invention, through the web crawler technology, the hot media information such as the popularity ranking, hot topics and hot celebrity dynamics of the whole network is monitored in real time, and then the relevant hot metadata in the missing and completed metadata is supplemented according to the hot media information, and a metadata framework is constructed to supplement the hot media that is not related to the missing and completed metadata to generate integrated metadata. At the same time, a hot media completion model is constructed, and the integrated metadata is imported into the hot media completion model. Through the hot media completion model, the integrated metadata is supplemented with AI to obtain the integrated and completed metadata.

[0063] Step S103: Establish a feature library of holidays and important social content, associate data features, and recommend specific media resources.

[0064] In an embodiment of the present invention, a feature library including holidays and important social contents such as family reunion dates, traditional custom dates and national commemoration dates is established, and then based on the feature library, data feature association is performed on the integrated and completed metadata, and feature association information is recorded. Then, based on the feature association information, specific media resources are recommended on relevant family reunion dates, traditional custom dates or national commemoration dates.

[0065] Step S104: collect user personalized information, build user behavior portraits and user personalized recommendation models, and generate personalized recommendation lists for users.

[0066] In an embodiment of the present invention, by collecting user personalized information including viewing history, collection records and search keyword information, and performing feature analysis on the user personalized information, a user behavior profile is constructed, and a user personalized recommendation model is constructed. By inputting the user behavior profile into the user personalized recommendation model, a personalized recommendation list is generated for the user through the user personalized recommendation model.

[0067] Furthermore, Figure 2 The structure diagram of the on-demand metadata information automatic completion system provided by an embodiment of the present invention is shown.

[0068] Specifically, in a preferred embodiment provided by the present invention, a system for automatically completing on-demand metadata information includes:

[0069] The stock media asset completion module 101 is used to collect and preprocess the existing metadata information of the stock media assets, build a stock media asset completion model, complete missing information, and obtain missing completion metadata.

[0070] In an embodiment of the present invention, the stock media asset completion module 101 collects metadata information such as the existing basic name, duration, and release time of the stock media assets, and then performs data cleaning on the metadata information to remove erroneous data and data in non-standard formats in the metadata information to obtain standard data. It also uses preset stock media asset metadata samples, sets hyperparameters such as learning rate and number of iterations, and performs model training processing on the convolutional neural network or the recurrent neural network until the model converges to generate a stock media asset completion model. The standard data is then input into the stock media asset completion model to predict and complete missing information, and obtain missing completion metadata with basic name, duration, release time, director, actor, and detailed type.

[0071] The hot media resource completion module 102 is used to monitor the hot media resource information of the entire network in real time, build a hot media resource completion model, and perform data integration and AI-assisted completion with the missing completion metadata to obtain integrated completion metadata.

[0072] In an embodiment of the present invention, the hot media resource completion module 102 uses web crawler technology to monitor the hot media resource information such as the popularity ranking, hot topics and hot celebrity dynamics of the entire network in real time, and then supplements the relevant hot media metadata in the missing completion metadata based on the hot media resource information, and constructs a metadata framework to supplement the hot media resources that are not related to the missing completion metadata to generate integrated metadata. At the same time, a hot media resource completion model is constructed, and the integrated metadata is imported into the hot media resource completion model. Through the hot media resource completion model, the integrated metadata is supplemented with AI to obtain the integrated completion metadata.

[0073] The specific media resource recommendation module 103 is used to establish a feature library of holidays and important social content, perform data feature association, and recommend specific media resources.

[0074] In an embodiment of the present invention, the specific media resource recommendation module 103 establishes a feature library including holidays and important social contents such as family reunion dates, traditional custom dates and national commemorative dates, and then, based on the feature library, associates data features of the integrated and completed metadata, records feature association information, and then, based on the feature association information, recommends specific media resources on relevant family reunion dates, traditional custom dates or national commemorative dates.

[0075] The user personalized recommendation module 104 is used to collect user personalized information, build user behavior portraits and user personalized recommendation models, and generate personalized recommendation lists for users.

[0076] In an embodiment of the present invention, the user personalized recommendation module 104 collects user personalized information including viewing history, collection records, and search keyword information, performs feature analysis on the user personalized information, constructs a user behavior profile, and constructs a user personalized recommendation model. By inputting the user behavior profile into the user personalized recommendation model, a personalized recommendation list is generated for the user through the user personalized recommendation model.

[0077] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0078] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0079] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims

1. A method for automatically completing on-demand metadata information, characterized in that: The method specifically comprises the following steps: Collect and preprocess the existing metadata information of stock media assets, build a stock media asset completion model, complete missing information, and obtain missing completion metadata; Monitor the hot media information of the entire network in real time, build a hot media completion model, and perform data integration and AI-assisted completion with the missing completion metadata to obtain integrated completion metadata; Establish a feature database of holidays and important social content, associate data features, and recommend specific media resources; Collect user personalized information, build user behavior portraits and user personalized recommendation models, and generate personalized recommendation lists for users.

2. The method for automatically completing on-demand metadata information according to claim 1, characterized in that: The data collection and preprocessing of the existing metadata information of the stock media assets, the construction of the stock media asset completion model, the completion of missing information, and the acquisition of missing completion metadata specifically include the following steps: Collect existing metadata information of stock media assets; Cleaning the metadata information to remove erroneous data and data in irregular formats to obtain standard data; Build a model to supplement existing media resources; The standard data is input into the stock media asset completion model to predict and complete missing information, thereby obtaining missing completion metadata.

3. The method for automatically completing on-demand metadata information according to claim 2, characterized in that: In constructing the existing media asset completion model, a preset existing media asset metadata sample is used, and hyperparameters are set to perform model training processing on a convolutional neural network or a recurrent neural network until the model converges to generate an existing media asset completion model, wherein the hyperparameters include a learning rate and a number of iterations.

4. The method for automatically completing on-demand metadata information according to claim 2, characterized in that: The metadata information includes a basic name, duration and release time; the missing completion metadata includes a basic name, duration, release time, director, actor and detailed type.

5. The method for automatically completing on-demand metadata information according to claim 1, characterized in that: The real-time monitoring of hot media information of the entire network, building a hot media completion model, and performing data integration and AI-assisted completion with the missing completion metadata to obtain the integrated completion metadata specifically include the following steps: Through web crawler technology, real-time monitoring of hot media information on the entire network; Integrate the hot media asset information with the missing and completed metadata to obtain integrated metadata; Construct a hot media resource completion model; The integrated metadata is completed with AI assistance through the hot media resource completion model to obtain integrated completed metadata.

6. The method for automatically completing on-demand metadata information according to claim 5, characterized in that: The hot media information includes popularity rankings, hot topics and hot celebrity dynamics.

7. The method for automatically completing on-demand metadata information according to claim 5, characterized in that: The step of integrating the hot media asset information with the missing and completed metadata to obtain the integrated metadata specifically comprises the following steps: According to the hot media asset information, supplement the relevant hot metadata in the missing and completed metadata; According to the hot media asset information, a metadata framework is constructed to supplement the hot media assets that are not related to the missing and completed metadata; Generates integration metadata.

8. The method for automatically completing on-demand metadata information according to claim 1, characterized in that: The steps of establishing a feature library of holidays and important social content, associating data features, and recommending specific media resources specifically include the following steps: Establishing a feature database of holidays and important social contents, wherein the feature database includes family reunion dates, traditional custom dates and national memorial dates; Based on the feature library, data feature association is performed on the integrated and completed metadata, and feature association information is recorded; Specific media resources are recommended based on the feature association information.

9. The method for automatically completing on-demand metadata information according to claim 1, characterized in that: The collecting of user personalized information, building of user behavior profiles and user personalized recommendation models, and generating personalized recommendation lists for users specifically include the following steps: Collect user personalized information, including viewing history, favorites, and search keyword information; Building a user behavior profile based on the user personalized information; Build a user personalized recommendation model; The user behavior profile is input into the user personalized recommendation model to generate a personalized recommendation list for the user.

10. A system for automatically completing metadata information on demand, characterized in that: The system includes a stock media resource supplement module, a hot media resource supplement module, a specific media resource recommendation module and a user personalized recommendation module, wherein: The stock media asset completion module is used to collect and preprocess the existing metadata information of the stock media assets, build a stock media asset completion model, complete the missing information, and obtain the missing completion metadata; A hot media resource completion module is used to monitor the hot media resource information of the entire network in real time, build a hot media resource completion model, and perform data integration and AI-assisted completion with the missing completion metadata to obtain integrated completion metadata; The specific media resource recommendation module is used to build a feature library of holidays and important social content, associate data features, and recommend specific media resources; The user personalized recommendation module is used to collect user personalized information, build user behavior portraits and user personalized recommendation models, and generate personalized recommendation lists for users.

Images