CN117972137A

CN117972137A - Intelligent supervision and analysis method, system and storage medium for video data

Info

Publication number: CN117972137A
Application number: CN202410386980.7A
Authority: CN
Inventors: 赵鑫; 隋阳; 岳平安
Original assignee: Shenzhen Zhishang Information Technology Co ltd
Current assignee: Shenzhen Zhishang Information Technology Co ltd
Priority date: 2024-04-01
Filing date: 2024-04-01
Publication date: 2024-05-03
Anticipated expiration: 2044-04-01
Also published as: CN117972137B

Abstract

The invention discloses an intelligent supervision and analysis method, system and storage medium for video data. Performing user behavior, interest characteristics and analysis according to the interaction data and the user basic data through a film and television platform, clustering and grouping the user characteristic data based on a K-means clustering algorithm, recommending contents to the users of the test group based on a plurality of recommendation strategies in a recommendation period, evaluating recommendation feedback of the users of the test group through the film and television platform, and screening out an optimal recommendation strategy; and monitoring the user terminal in real time, collecting user feedback data, respectively carrying out recommendation effect evaluation calculation on N user cluster groups through the user feedback data to obtain N evaluation values, carrying out weighted average calculation on the N evaluation values based on the number of users in the user cluster groups, and obtaining the overall recommendation evaluation value. According to the invention, the recommendation data can be accurately and rapidly analyzed, so that the film and television requirements of huge and complex users can be met.

Description

Intelligent supervision and analysis method, system and storage medium for video data

Technical Field

The invention relates to the field of film and television data analysis, in particular to an intelligent supervision and analysis method, system and storage medium for film and television data.

Background

With the rapid development of the internet, movie data is explosively increased, and users' demands for video contents are increasingly diversified. The traditional film and television data supervision and analysis method mainly relies on manual auditing and analysis, has low efficiency and cannot meet the requirement of large-scale data processing. Meanwhile, due to the fact that the traditional technology is adopted, the data recommendation effect aiming at the users of the same category is often not ideal, and accurate user feature analysis and data recommendation are difficult to achieve for the users, so that how to achieve supervision analysis and classification of intelligent video data and how to provide accurate content recommendation is an important problem to be solved at present.

Disclosure of Invention

The invention overcomes the defects of the prior art and provides an intelligent supervision and analysis method, system and storage medium for video data.

The first aspect of the invention provides an intelligent supervision and analysis method for video data, which comprises the following steps:

collecting interaction data and user basic data of all users currently through a film and television platform in a preset time period;

analyzing and extracting user behavior characteristics and interest characteristics according to the interaction data and user basic data, obtaining user characteristic data, clustering and grouping the user characteristic data based on a K-means clustering algorithm, and classifying users into N user clustering groups based on a clustering result;

Selecting all user characteristic data in a user cluster group, and carrying out content recommendation strategy analysis based on all user characteristic data to obtain various recommendation strategies;

Dividing two groups of users in a user cluster group, correspondingly being a test group user and a non-test group user, recommending the content of the test group user based on a plurality of recommendation strategies in one recommendation period, evaluating recommendation feedback of the test group user through a film and television platform, screening out an optimal recommendation strategy, and recommending data of the user cluster group according to the optimal recommendation strategy in the next recommendation period;

Analyzing all user cluster groups, and screening N optimal recommendation strategies to be applied to user terminals corresponding to each user cluster group;

And in the next recommendation period, monitoring the user terminal in real time through a film and television platform, collecting user feedback data, respectively carrying out recommendation effect evaluation calculation on N user cluster groups through the user feedback data to obtain N evaluation values, carrying out weighted average calculation on the N evaluation values based on the number of users in the user cluster groups, and obtaining the overall recommendation evaluation value.

In this scheme, in a preset time period, collect interaction data and user basic data of all users currently through the movie platform, specifically:

the user terminal is connected with the Internet through the film and television platform and interacted with the data through the network;

collecting interactive data of all current users in a preset time period, wherein the interactive data comprise video content on demand of the users, search content, advertisement click content and user login platform use data;

retrieving user information data of all current users from a system database through a film and television platform, and integrating the user information data into user basic data;

The user basic data comprises user account information, historical video records and historical login data.

In this scheme, the analysis and extraction of user behavior features and interest features are performed according to the interaction data and user basic data, and user feature data is obtained, the user feature data is clustered and grouped based on a K-means clustering algorithm, and the users are classified into N user cluster groups based on a clustering result, specifically:

analyzing and extracting user behavior characteristics and user interest characteristics of each user according to the interaction data and the user basic data to obtain behavior characteristic data and interest characteristic data;

Carrying out data integration on the behavior characteristic data and the interest characteristic data of each user to form user characteristic data;

setting N clustering groups based on a K-means clustering algorithm, and clustering and grouping by taking user characteristic data as clustering sample data to obtain a clustering and grouping result;

And carrying out corresponding user grouping division according to the clustering grouping result, and obtaining N user clustering groups, wherein each user clustering group comprises a plurality of users.

In this scheme, all user feature data in a user cluster group are selected, and content recommendation policy analysis is performed based on all user feature data to obtain multiple recommendation policies, specifically:

selecting one of N user cluster groups, marking the selected cluster group as a current cluster group, and integrating user characteristic data of all users in the current cluster group to form user characteristic big data;

performing data integration, data cleaning, data redundancy removal and data outlier removal pretreatment on the user characteristic big data;

Carrying out data recommendation analysis based on collaborative filtering algorithm on the preprocessed user characteristic big data, and obtaining corresponding recommended video data;

carrying out user habit behavior analysis on the preprocessed user characteristic big data, and obtaining various recommended time strategies;

Classifying the recommended video data based on video content to form various recommended video tag information;

And combining the plurality of recommended time strategies with the plurality of recommended video tag information to form a plurality of recommended strategies.

In this scheme, two groups of users are divided from one user cluster group, the two groups of users are corresponding to a test group user and a non-test group user, content recommendation is performed on the test group user based on multiple recommendation strategies in one recommendation period, recommendation feedback of the test group user is evaluated through a video platform, and an optimal recommendation strategy is screened out, and in the next recommendation period, data recommendation is performed on one user cluster group according to the optimal recommendation strategy, specifically:

dividing two groups of users from the current cluster group according to a preset proportion, and marking the two groups of users as test group users and non-test group users respectively;

In a recommendation period, applying a plurality of recommendation strategies to test group users, wherein each user in the test group users corresponds to one recommendation strategy, and recording application association information of the recommendation strategies and the users;

In a recommendation period, collecting feedback data of users of a test group in real time through a video platform, wherein the feedback data comprises residence time, click frequency and jump information of the users watching the recommended video data;

according to the feedback data, carrying out recommended satisfaction evaluation on the users of the test group, and screening out the optimal recommended strategy with highest satisfaction degree by combining with application associated information;

and in the next recommendation period, recommending the real-time movie and television data to the users of the current cluster group based on the optimal recommendation strategy.

In this scheme, the analysis of all user cluster groups and the screening of N optimal recommendation strategies applied to the user terminals corresponding to each user cluster group specifically includes:

in a recommendation period, analyzing all user cluster groups, and screening N optimal recommendation strategies;

Each user cluster group corresponds to an optimal recommendation strategy;

And applying N optimal recommendation strategies to the user terminals in each user cluster group.

In this scheme, in the next recommendation period, the user terminal is monitored in real time through the movie platform and user feedback data is collected, recommendation effect evaluation calculation is performed on N user cluster groups through the user feedback data respectively to obtain N evaluation values, and based on the number of users in the user cluster groups, weighted average calculation is performed on the N evaluation values to obtain an overall recommendation evaluation value, which specifically includes:

in the next recommendation period, monitoring the user terminal in real time through the video platform, collecting user feedback data and marking the user feedback data as second feedback data;

k users are selected from each user cluster group to serve as evaluation users, terminal picture interception is carried out on the evaluation users in the next recommendation period time period, and intercepted image data are obtained;

taking the evaluation users in each user cluster group as analysis units, carrying out picture content recognition by intercepting image data to obtain picture recognition results, carrying out content matching analysis on the picture recognition results and the corresponding optimal recommendation strategies, and obtaining N recommendation matching degrees;

respectively carrying out recommendation effect evaluation calculation analysis on the N user clustering groups according to the user feedback data, and obtaining N evaluation values;

and carrying out weighted average calculation on all the user clustering groups based on the recommendation matching degree and the evaluation value, wherein the weight is based on the number of users in each user clustering group, and obtaining the overall recommendation evaluation value.

In this scheme, the overall recommendation evaluation value is obtained, further including:

judging whether the overall recommendation evaluation value is smaller than a preset threshold value or not in the current recommendation period;

If the real-time interaction data is smaller than the real-time interaction data, real-time interaction data is collected, and secondary user feature analysis and secondary user clustering grouping are performed based on the real-time interaction data;

Based on the secondary user clustering grouping result, analyzing a plurality of corresponding recommendation strategies, and performing secondary test and feedback data analysis on the user clustering group to obtain a second optimal recommendation strategy;

updating the original optimal recommendation strategy by the obtained second optimal recommendation strategy, and updating corresponding recommendation data at the same time;

and if the recommendation strategy is not smaller than the preset recommendation strategy, applying the optimal recommendation strategy based on the current recommendation period to the next recommendation period.

The second aspect of the present invention also provides an intelligent supervision and analysis system for video data, which comprises: the device comprises a memory and a processor, wherein the memory comprises a film and television data intelligent supervision and analysis program, and the film and television data intelligent supervision and analysis program realizes the following steps when being executed by the processor:

The third aspect of the present invention also provides a computer readable storage medium, where the computer readable storage medium includes a video data intelligent supervision and analysis program, where the video data intelligent supervision and analysis program, when executed by a processor, implements the steps of the video data intelligent supervision and analysis method according to any one of the above.

Drawings

FIG. 1 shows a flow chart of an intelligent supervision and analysis method for video data according to the invention;

FIG. 2 shows a flowchart for acquiring a user cluster set in accordance with the present invention;

fig. 3 shows a block diagram of an intelligent supervisory analysis system for video data according to the present invention.

Detailed Description

In order that the above-recited objects, features and advantages of the present application will be more clearly understood, a more particular description of the application will be rendered by reference to the appended drawings and appended detailed description. It should be noted that, without conflict, the embodiments of the present application and features in the embodiments may be combined with each other.

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present invention, but the present invention may be practiced in other ways than those described herein, and therefore the scope of the present invention is not limited to the specific embodiments disclosed below.

Fig. 1 shows a flowchart of an intelligent supervision and analysis method for video data according to the present invention.

As shown in fig. 1, the first aspect of the present invention provides an intelligent supervision and analysis method for video data, which includes:

s102, collecting interaction data and user basic data of all current users through a film and television platform in a preset time period;

s104, analyzing and extracting user behavior characteristics and interest characteristics according to the interaction data and user basic data, obtaining user characteristic data, clustering and grouping the user characteristic data based on a K-means clustering algorithm, and classifying users into N user clustering groups based on a clustering result;

S106, selecting all user characteristic data in a user cluster group, and carrying out content recommendation strategy analysis based on all user characteristic data to obtain various recommendation strategies;

S108, dividing two groups of users in a user cluster group, correspondingly to test group users and non-test group users, recommending contents to the test group users based on multiple recommendation strategies in a recommendation period, evaluating recommendation feedback of the test group users through a film and television platform, screening out an optimal recommendation strategy, and recommending data to the user cluster group according to the optimal recommendation strategy in the next recommendation period;

s110, analyzing all user cluster groups, and screening N optimal recommendation strategies to be applied to user terminals corresponding to each user cluster group;

And S112, in the next recommendation period, monitoring the user terminal in real time through a film and television platform, collecting user feedback data, respectively carrying out recommendation effect evaluation calculation on N user cluster groups through the user feedback data to obtain N evaluation values, carrying out weighted average calculation on the N evaluation values based on the number of users in the user cluster groups, and obtaining the overall recommendation evaluation value.

According to the embodiment of the invention, in a preset time period, the interactive data and the user basic data of all the current users are collected through the film and television platform, specifically:

The user login platform use data includes login platform time, use time, video data transmission quantity and other platform record data, and is used for analyzing user behavior characteristics and the like.

FIG. 2 shows a flow chart for acquiring a user cluster set in accordance with the present invention.

According to the embodiment of the invention, the analysis and extraction of the user behavior characteristics and the interest characteristics are carried out according to the interaction data and the user basic data, the user characteristic data is obtained, the user characteristic data is clustered and grouped based on a K-means clustering algorithm, and the users are divided into N user clustering groups based on a clustering result, specifically:

S202, analyzing and extracting user behavior characteristics and user interest characteristics of each user according to the interaction data and user basic data to obtain behavior characteristic data and interest characteristic data;

S204, carrying out data integration on the behavior characteristic data and the interest characteristic data of each user to form user characteristic data;

s206, setting N clustering groups based on a K-means clustering algorithm, and clustering and grouping by taking user characteristic data as clustering sample data to obtain a clustering and grouping result;

S208, carrying out corresponding user grouping division according to the clustering grouping result, and obtaining N user clustering groups, wherein each user clustering group comprises a plurality of users.

The behavior characteristic data specifically refer to login behavior of a user, operation behavior on a video platform, use time, use frequency and the like, and the interest characteristic data mainly comprises characteristic data such as user interest video tags, user interest advertisement content, user interest habits and the like. Each user corresponds to a piece of user characteristic data. The clustering grouping result comprises N groups of data, each group of data comprises one or more user characteristic data, and the clustering grouping result is further used for mapping and dividing user groups. In each user clustering group, the user behavior characteristics and the interest characteristics in the same group have the characteristic of high similarity, and based on the K-means clustering grouping, the users with the same characteristics can be effectively grouped, so that large-scale user data analysis is reduced, and accurate and rapid user data recommendation is realized.

According to the embodiment of the invention, all user characteristic data in one user cluster group are selected, and content recommendation strategy analysis is performed based on all the user characteristic data to obtain various recommendation strategies, specifically:

It should be noted that, the recommended time policy includes a time node of the recommended data, a duration of the recommendation process, and the like, and is obtained based on the analysis of the relevant user features, and the recommended movie tag information is a corresponding interest content classification tag, such as a movie classification tag of military, news, entertainment, sports, travel, and the like.

One recommendation policy includes one recommendation video tag information and one recommendation time policy, for example, if there are N recommendation time policies and M recommendation video tag information, n×m recommendation policies may be formed in total. And randomly applying N multiplied by M recommendation strategies to the test user during the subsequent test, so that each recommendation strategy has application corresponding to the test user as far as possible.

According to the embodiment of the invention, two groups of users are divided in one user cluster group, the two groups of users are corresponding to a test group user and a non-test group user, content recommendation is carried out on the test group user based on a plurality of recommendation strategies in one recommendation period, recommendation feedback of the test group user is evaluated through a video platform, an optimal recommendation strategy is screened out, and data recommendation is carried out on the one user cluster group according to the optimal recommendation strategy in the next recommendation period, specifically:

It should be noted that, the preset proportion is generally that the test group users occupy a relatively small proportion, and users can be divided equally based on actual application conditions. And the jump information is interaction record information in the process of clicking and jumping other content information after the user views the recommended content, and is used for judging and analyzing the effect evaluation of the user on the recommended data.

According to the embodiment of the invention, all user cluster groups are analyzed, N optimal recommendation strategies are screened out and applied to the user terminals corresponding to each user cluster group, and the method specifically comprises the following steps:

Each user cluster group corresponds to an optimal recommendation strategy;

It should be noted that, there is often a large difference in the optimal recommendation strategy between different user cluster groups, and the difference is determined by the difference of the user features between the cluster groups.

According to the embodiment of the invention, in the next recommendation period, the user terminal is monitored in real time through the video platform and user feedback data is collected, recommendation effect evaluation calculation is respectively carried out on N user cluster groups through the user feedback data to obtain N evaluation values, weighted average calculation is carried out on the N evaluation values based on the number of users in the user cluster groups, and an overall recommendation evaluation value is obtained, specifically:

It should be noted that, each user cluster group corresponds to K evaluation users, and n×k evaluation users in total, generally, K has a smaller value, and is used for randomly selecting user evaluation, so as to reduce the calculation amount of the system. And in the process of carrying out content matching analysis on the picture identification result and the corresponding optimal recommendation strategy and obtaining the recommendation matching degree, specifically, carrying out matching analysis on the picture identification result and the corresponding recommendation film and television label information in the optimal recommendation strategy, wherein the higher the recommendation matching degree is, the higher the matching degree between the user browsing content and the corresponding recommendation strategy is, and the better the recommendation strategy effect is. The evaluation value is obtained based on feedback data analysis, and is mainly based on information analysis recommendation effects such as user browsing behavior characteristic content and the like. The overall recommendation evaluation value reflects the recommendation effect of the whole user group, and can scientifically and accurately analyze the quick recommendation effect of each recommendation period or within a certain time period through the overall recommendation evaluation value and further judge whether a recommendation strategy needs to be replaced or not so as to dynamically update recommendation data.

And taking each user cluster group as an analysis unit, and carrying out mean value calculation on the corresponding recommended matching degree and the evaluation value, wherein the calculation formula is as follows: Wherein N is the number of user cluster groups, For the weight of the ith user cluster group, the weight is replaced by the number of users of the user cluster group, namely, the greater the number of users is, the greater the weight is,/>For the preset correction coefficient, F is the overall recommended evaluation value,/>And respectively recommending matching degree and evaluation value of the ith user cluster group.

According to an embodiment of the present invention, the obtaining the overall recommendation evaluation value further includes:

It is worth mentioning that, in the long-time movie platform operation process, the change of the hot spot current events and hot movies and television also can change the interest characteristics of the corresponding users, so that the invention uses the overall recommendation evaluation value as a measurement standard, when the overall recommendation evaluation value is low, the invention dynamically analyzes the secondary recommendation data of the users in real time and dynamically changes the recommendation strategy, thereby being capable of adapting to the complicated and changed movie platform environment of the users and improving the instantaneity and the accuracy of the recommendation data.

In addition, the recommendation flow method with high efficiency and low system consumption can be provided under a large-scale user platform, and the recommendation flow method has a better effect in user recommendation analysis with high user quantity and high complexity. It is worth mentioning that after the video platform is used for a period of time, the user may have interest characteristics or behavior characteristics changed, which may be subject to real-time change of current events, so that if the recommendation strategy is fixed, it is difficult to dynamically adjust the recommendation data and adapt the system to the user, therefore, the invention performs feedback analysis based on the clustered users, evaluates the recommendation effect of the integrity by weighting, and then performs threshold comparison based on the weighted evaluation value, determines whether the secondary user feature data acquisition analysis and secondary clustering grouping are needed, determines whether the recommendation strategy update is needed, thereby realizing the purpose of dynamically adjusting the user recommendation data, and the weighted index performs comprehensive evaluation based on the clustered users, so as to realize scientific and accurate platform integrity recommendation effect analysis,

It is worth mentioning that under the long-term operation of the video platform, the behavior and interest characteristics of each group of users in the clustering group have certain changes, the changes can reflect corresponding feedback data, if the calculation analysis of the feedback data and the real-time analysis of the browsing data of the users are carried out on each user, the calculation amount is often large and is difficult to realize due to the limitation of hardware conditions, therefore, the invention can accurately and quickly obtain the recommended data by carrying out grouping analysis of the recommended strategies on the form of the user clustering groups and selecting the optimal strategies through certain test feedback data so as to meet the video requirements of huge and complex users.

According to an embodiment of the present invention, further comprising:

under a recommendation period, acquiring real-time interaction data of all users, and dividing the real-time interaction data into N user interaction data based on N user clustering groups;

Selecting one user cluster group as a sample cluster group, extracting and analyzing film and television contents in user interaction data in the sample cluster group, and classifying based on film and television content labels to obtain a plurality of real-time film and television labels;

In the classifying process, carrying out statistical analysis and heat evaluation on data in three dimensions of corresponding browsing time, browsing frequency and browsing click times in different types of film and television data, and obtaining the browsing frequency of each real-time film and television label;

Dividing K continuous recommendation periods in a latest preset time period, and carrying out data integration and data serialization on the real-time video tags of the sample cluster groups and the corresponding browsing frequencies by taking time as a unit in the K continuous recommendation periods to form user serialization data;

based on an LSTM prediction algorithm, carrying out data prediction by taking user serialized data as reference data, and obtaining predicted sequence data;

Analyzing the predicted sequence data, analyzing a real-time video tag with the browsing frequency having an ascending trend, and marking the real-time video tag to obtain predicted video tag information;

analyzing all N user cluster groups, obtaining N predicted video label information, screening video labels with highest occurrence frequency based on the N predicted video label information, and marking the video labels as high-frequency video labels;

and generating additional recommended content by combining a video platform based on the high-frequency video tag, and importing the additional recommended content into an optimal recommendation strategy.

The predicted video tag information includes predicted browsing frequency information of the predicted video tag. And the video labels with highest frequency of occurrence are screened out and analyzed based on predicted browsing frequency information.

In addition, in K continuous recommendation periods, each period is provided with a plurality of corresponding real-time video tags and a plurality of browsing frequencies, the real-time video tags and browsing frequency data of each period are serialized based on time continuity to form user serialized data, in the plurality of continuous periods, the types and the numbers of the video tags of the users possibly change to a certain extent, the core change is analyzed to be the browsing frequency change of different tags, through serializing the data, the prediction analysis of interest data of the users of the sample cluster group can be performed based on time dimension, and finally, corresponding additional recommendation data is recommended based on prediction results. It is worth mentioning that in the whole user group, because of the influence of real-time hot spots and fresh hot video data, the interest content of different types of users is likely to have content with consistent trend, the content generally occupies lower content but has higher user applicability and universality, and generally has higher real-time analysis requirements, and the content is difficult to mine and recommend due to the traditional recommendation analysis scheme.

It should be noted that, the browsing frequency is a heat value, which is proportional to the browsing time, the browsing frequency and the number of browsing clicks, and reflects the heat of a certain video tag content, so as to reflect which type of video content has a higher interest or higher browsing frequency by the user, and further predict the interest trend of the user. Each user cluster group corresponds to a plurality of real-time video labels and a plurality of browsing frequencies.

The second aspect of the present invention also provides an intelligent supervisory analysis system 3 for video data, which comprises: the memory 31 and the processor 32, wherein the memory comprises a video data intelligent supervision and analysis program, and the video data intelligent supervision and analysis program realizes the following steps when being executed by the processor:

Each user cluster group corresponds to an optimal recommendation strategy;

And taking each user cluster group as an analysis unit, and carrying out mean value calculation on the corresponding recommended matching degree and the evaluation value, wherein the calculation formula is as follows: Wherein N is the number of user cluster groups, For the weight of the ith user cluster group, the weight is replaced by the number of users of the user cluster group, namely, the greater the number of users is, the greater the weight is,/>For the preset correction coefficient, F is the overall recommended evaluation value,/>And respectively recommending matching degree and evaluation value of the ith user cluster group. According to an embodiment of the present invention, the obtaining the overall recommendation evaluation value further includes:

According to an embodiment of the present invention, further comprising:

In the several embodiments provided by the present application, it should be understood that the disclosed apparatus and method may be implemented in other ways. The above described device embodiments are only illustrative, e.g. the division of the units is only one logical function division, and there may be other divisions in practice, such as: multiple units or components may be combined or may be integrated into another system, or some features may be omitted, or not performed. In addition, the various components shown or discussed may be coupled or directly coupled or communicatively coupled to each other via some interface, whether indirectly coupled or communicatively coupled to devices or units, whether electrically, mechanically, or otherwise.

The units described above as separate components may or may not be physically separate, and components shown as units may or may not be physical units; can be located in one place or distributed to a plurality of network units; some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated in one processing unit, or each unit may be separately used as one unit, or two or more units may be integrated in one unit; the integrated units may be implemented in hardware or in hardware plus software functional units.

Those of ordinary skill in the art will appreciate that: all or part of the steps for implementing the above method embodiments may be implemented by hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, where the program, when executed, performs steps including the above method embodiments; and the aforementioned storage medium includes: a mobile storage device, a Read-Only Memory (ROM), a random access Memory (RAM, random Access Memory), a magnetic disk or optical disk, or the like, which can store program codes.

Or the above-described integrated units of the invention may be stored in a computer-readable storage medium if implemented in the form of software functional modules and sold or used as separate products. Based on such understanding, the technical solutions of the embodiments of the present invention may be embodied in essence or a part contributing to the prior art in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the methods described in the embodiments of the present invention. And the aforementioned storage medium includes: a removable storage device, ROM, RAM, magnetic or optical disk, or other medium capable of storing program code.

The foregoing is merely illustrative of the present invention, and the present invention is not limited thereto, and any person skilled in the art will readily recognize that variations or substitutions are within the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. An intelligent supervision and analysis method for video data is characterized by comprising the following steps:

2. The method for intelligently supervising and analyzing video data according to claim 1, wherein the step of collecting the interaction data and the user base data of all the users currently through the video platform within a preset time period is specifically as follows:

3. The intelligent supervision and analysis method of video data according to claim 2, wherein the analyzing and extracting of the user behavior feature and the interest feature are performed according to the interaction data and the user basic data, and user feature data is obtained, the user feature data is clustered and grouped based on a K-means clustering algorithm, and the users are classified into N user cluster groups based on a clustering result, specifically:

4. The method for intelligently supervising and analyzing video data according to claim 3, wherein the selecting all user characteristic data in a user cluster group and analyzing content recommendation strategies based on the all user characteristic data, and obtaining a plurality of recommendation strategies comprises the following specific steps:

5. The intelligent supervision and analysis method for video data according to claim 4, wherein two groups of users are divided in a user cluster group, the users are corresponding to a test group user and a non-test group user, content recommendation is performed on the test group user based on a plurality of recommendation strategies in a recommendation period, recommendation feedback of the test group user is evaluated through a video platform, an optimal recommendation strategy is screened out, and data recommendation is performed on the user cluster group according to the optimal recommendation strategy in a next recommendation period, specifically:

6. The method for intelligently supervising and analyzing video data according to claim 5, wherein the analyzing all user cluster groups and screening out N optimal recommendation strategies is applied to user terminals corresponding to each user cluster group comprises:

Each user cluster group corresponds to an optimal recommendation strategy;

7. The method for intelligently supervising and analyzing video data according to claim 6, wherein in the next recommendation period, the user terminal is supervised in real time and user feedback data is collected through the video platform, recommendation effect evaluation calculation is performed on N user cluster groups through the user feedback data respectively to obtain N evaluation values, weighted average calculation is performed on the N evaluation values based on the number of users in the user cluster groups, and an overall recommendation evaluation value is obtained, specifically:

8. The method for intelligently supervising and analyzing video data according to claim 7, wherein the obtaining of the overall recommendation evaluation value further comprises:

9. An intelligent supervisory analysis system for video data, which is characterized in that the system comprises: the device comprises a memory and a processor, wherein the memory comprises a film and television data intelligent supervision and analysis program, and the film and television data intelligent supervision and analysis program realizes the following steps when being executed by the processor:

10. A computer-readable storage medium, wherein the computer-readable storage medium includes therein a video data intelligent supervisory analysis program, which when executed by a processor, implements the steps of the video data intelligent supervisory analysis method according to any one of claims 1 to 8.