Social platform information spreading intelligent management system based on man-machine symbiosis

Abstract

The invention belongs to the field of social platform information spreading management, and relates to a social platform information spreading intelligent management system based on man-machine symbiosis, which is characterized in that according to an information credibility priority principle and an information diversity optimization principle, primary intervention operation and secondary intervention operation are respectively carried out on a currently investigated video set; the first intervention operation ensures that the target video platform obtains sufficient exposure opportunities for high-quality video contents of a specified user, the overall content quality guarantee of the platform for video pushing of the user is improved, and the second intervention operation focuses on reducing invalid browsing with repeated and similar contents and pays attention to the comprehensiveness of the contents; the method is advantaged in that a user is prevented from falling into an information cocoon room, then through objective and accurate evaluation of the actual effect of the video browsing sorting intervention measure, a scientific basis is provided for optimization and adjustment of follow-up platform pushing video browsing sorting intervention, and viscidity and loyalty of the user to the platform are greatly enhanced.

Description

Technical Field

[0001] The present invention belongs to the field of social platform information management, and relates to an intelligent management system for social platform information dissemination based on human-machine symbiosis. Background Art

[0002] As an important channel for information dissemination, the efficiency and effectiveness of social platforms in information dissemination directly affect the direction of social public opinion and public cognition. However, due to the complexity and diversity of information dissemination, information dissemination on social platforms is often affected by many factors, including the intention of the disseminator, the cognitive level of the audience, the authenticity of the information content, etc. Therefore, how to effectively manage the information dissemination on social platforms and improve the efficiency and effectiveness of information dissemination has become an urgent problem to be solved.

[0003] There are also some existing technologies that involve related solutions such as information dissemination management on social platforms. For example, China Patent Publication No. CN118840077A is a blockchain-based social platform content management system. It adjusts the content review priority based on user activity by evaluating user historical behavior data, identifies potential illegal content by comparing and analyzing the published content with the banned content library, and records the editing and sharing process of the content on the content blockchain. When illegal content appears, it quickly identifies and punishes key dissemination users, improves the management efficiency and accuracy of illegal behaviors, and creates a safer and healthier communication environment for users.

[0004] The Chinese patent publication number is CN114004709A, which is a monitoring method, device and computer-readable storage medium for information dissemination. The method obtains data on the user's participation in the information dissemination process from a social platform, preprocesses the data according to the time period of the dissemination, obtains the time interval matrix and relative position matrix in the information dissemination process from the preprocessed data, adopts a self-attention mechanism to obtain the user's temporal structure information and position structure information in the information dissemination process and fuses them, inputs the fused information into the information dissemination monitoring model, dynamically monitors the trajectory of the user's participation in information dissemination at the next time, helps to better extract the characteristic information of information dissemination, and monitors the information dissemination process.

[0005] The Chinese patent publication number CN104408659A discloses an evaluation method and system based on the information propagation model of social applications. The method analyzes and abstracts the information propagation process of social applications, establishes the information propagation model of social applications, decomposes the information propagation risk of social applications into various factors according to a hierarchical relationship, and aggregates and combines the various factors according to a hierarchical relationship based on the mutual relationship between the various factors to generate a false information propagation risk evaluation model. The false information propagation threat faced by the social application is evaluated based on the risk points in the risk point layer of the false information propagation risk evaluation model as the basic original data.

[0006] However, the above scheme still involves limitations that are common in existing technologies. Specifically, after a series of processing such as the identification of illegal virtual risk content in the above scheme, existing technologies usually generate a list of videos to be pushed or spread for social platform users that fit their browsing behavior habits. However, existing technologies lack a browsing ranking intervention mechanism for videos that social platform users intend to push or spread. On the one hand, it may result in high-quality content not being given priority for display opportunities, while low-quality video content excessively occupies the user's browsing window, causing the overall quality of the platform content ecology to decline, thereby affecting user stickiness and activity.

[0007] On the other hand, users may frequently encounter video push notifications with similar or repeated content, resulting in serious homogeneity of the platform's video push content, making it easy for users to fall into the dilemma of information cocoons. It may also cause users' aesthetic fatigue, thereby weakening their viewing experience.

[0008] Furthermore, the inadequacy of the browsing sorting control mechanism also means that the videos to be pushed have not undergone a secondary review and retrieval process, and thus cannot provide users with a higher level of authenticity and reliability assurance of video information. Summary of the invention

[0009] In view of this, in order to solve the problems raised in the above background technology, an intelligent management system for information dissemination on a social platform based on human-machine symbiosis is proposed.

[0010] The purpose of the present invention can be achieved through the following technical solutions: The present invention provides an intelligent management system for information dissemination on a social platform based on human-computer symbiosis, including: a module for extracting a set of videos to be pushed, a primary intervention module for video browsing and sorting, a secondary intervention module for video browsing and sorting, an intervention effect evaluation module and a cloud database.

[0011] The proposed video set extraction module is connected to the video browsing sorting primary intervention module, the video browsing sorting primary intervention module is connected to the video browsing sorting secondary intervention module, the video browsing sorting secondary intervention module is connected to the intervention effect evaluation module, and the cloud database is connected to the video browsing sorting primary intervention module.

[0012] The module for extracting the video set to be pushed is used to extract the video set to be pushed by the target video platform for the specified user within a preset time period in the future on the same day, which is recorded as the current inspection video set.

[0013] The video browsing sorting one-time intervention module is used to analyze the publishing source reliability, publishing content credibility and publishing dissemination positive trend of each video in the current inspection video set according to the principle of information credibility priority, so as to intervene in the video browsing sorting of the current inspection video set.

[0014] The video browsing ranking secondary intervention module is used to perform secondary intervention on the video browsing ranking of the current inspection video set according to the information diversity optimization principle.

[0015] The intervention effect evaluation module is used to use the video browsing ranking after the secondary intervention of the current inspection video set as the actual video playback ranking, collect the actual browsing behavior data of the specified user in the future preset time period of the day, evaluate the intervention effect of the video browsing ranking of the current inspection video set and provide feedback.

[0016] The cloud database is used to store the target video platform's standardized sensitive vocabulary set and sensitive image element set, and to store the positive trending vocabulary set and non-positive trending vocabulary set corresponding to the user comment retrieval specification.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention analyzes the reliability of the publishing source, the credibility of the publishing content and the positive trend of the publishing and dissemination of each video in the current video set, and conducts an intervention operation on the video browsing order of the current video set according to the principle of information credibility priority. This not only ensures that high-quality content has sufficient exposure opportunities and improves users' trust and satisfaction with the platform, but also effectively identifies and intercepts non-high-quality content, thereby improving the overall content quality of the platform's video push to users.

[0018] (2) Based on the information diversity optimization principle, the present invention combines the two dimensions of video theme type and video browsing length to carry out secondary intervention operations on the video browsing ranking of the current video set. This not only reduces invalid browsing caused by users' lack of interest or repeated and similar content, and improves the user's overall browsing experience, but also focuses on the comprehensiveness of the content, avoiding users from being trapped in information cocoons, and greatly enhancing users' stickiness and loyalty to the platform.

[0019] (3) The present invention helps to more objectively and accurately evaluate the actual effects of video browsing ranking intervention measures by collecting actual browsing behavior data of designated users on the target video platform, and provides a scientific basis for the optimization and adjustment of subsequent platform push video browsing ranking intervention. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for describing the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0021] Figure 1 It is a schematic diagram of module connection of the present invention. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0023] See also Figure 1 As shown, the present invention provides an intelligent management system for information dissemination on a social platform based on human-machine symbiosis, including: a module for extracting a set of videos to be pushed, a module for first-time intervention in video browsing and sorting, a module for second-time intervention in video browsing and sorting, an intervention effect evaluation module and a cloud database.

[0024] The proposed video set extraction module is connected to the video browsing sorting primary intervention module, the video browsing sorting primary intervention module is connected to the video browsing sorting secondary intervention module, the video browsing sorting secondary intervention module is connected to the intervention effect evaluation module, and the cloud database is connected to the video browsing sorting primary intervention module.

[0025] The proposed video set extraction module is used to extract the proposed video set to be pushed by the target video platform for the designated user within a preset time period in the future on the same day, which is recorded as the current inspection video set.

[0026] The video browsing sorting one-time intervention module is used to analyze the publishing source reliability, publishing content credibility and publishing dissemination positive trend of each video in the current inspection video set according to the information credibility priority principle, so as to intervene in the video browsing sorting of the current inspection video set.

[0027] Specifically, the analysis of the reliability of the publishing source of each video in the current video set includes: collecting the publishing author information of each video in the current video set, including the official certification indicators of the target video platform, registration years, community participation contribution, number of violations and the severity of each violation, fan size, fan activity and fan retention rate, and analyzing the influence, recognition and risk of the publishing author information, so as to evaluate the basic credibility coefficient of each video publishing author in the current video set.

[0028] It should be noted that the specific process of analyzing the influence, recognition and risk of the publishing author information includes: extracting the official certification indicators, registration years and community participation contribution of the target video platform where the publishing author information is located, and performing ratio analysis with the official certification reference indicators, reference registration years and reference community participation contribution preset on the target video platform corresponding to the video publishing author recognition evaluation, and accumulating the calculated ratios to obtain the analysis result of the publishing author information recognition.

[0029] The fan size, fan activity and fan retention rate of the target video platform where the publishing author information is located are extracted, and ratio analysis is performed on them with the reference fan size, reference fan activity and reference fan retention rate preset on the target video platform corresponding to the video publishing author influence evaluation. The calculated ratios are accumulated to obtain the analytical result of the influence of the publishing author information.

[0030] Extract the number of violations of the target video platform where the author information is published and the severity of each violation , is the number of each violation, , according to the formula Analyze and obtain the analysis results of the risk of publishing author information, including They are respectively the reference number of violations preset in the target video platform and the reference severity of a single violation corresponding to the risk assessment of the video publisher.

[0031] It should be noted that the above-mentioned official certification status indicator takes a value of 0 or 1, where 0 indicates that the video publisher has not been officially certified, and 1 indicates that the video publisher has been officially certified. The official certification reference indicator value of the target video platform corresponding to the recognition evaluation of the video publisher is 1. The severity of each violation of the target video platform in the above-mentioned publisher information is the corresponding severity value of each type of violation penalty predefined by the target video platform, and the corresponding assessment is made through the type of violation penalty imposed on the publisher by the target video platform.

[0032] Retrieve the historical published content set of the authors who published each video in the currently examined video set, analyze the content information density, content feedback activity and content viewing comfort of each video in the historical published content set, and analyze the content quality coefficient of each video in the historical published content set, so as to evaluate the content creation quality coefficient of each video publisher in the currently examined video set.

[0033] It should be noted that the specific process of analyzing the content information density, content feedback activity and content viewing comfort of each video in the historically released content set includes: converting the audio content of each video in the historically released content set into text content through speech recognition technology, extracting each keyword in the video text content and identifying its corresponding involved fields, counting the number of text keywords and the number of involved fields of each video in the historically released content set, and performing ratio analysis with the preset reference text keyword number and the preset reference involved field number, respectively, and accumulating the results of each ratio analysis to obtain the content information density of each video in the historically released content set.

[0034] Collect the number of likes, shares, and comments on each video in the historical content set as of the day, perform ratio analysis with the preset reference video likes, video shares, and video comments, and add up the results of each ratio analysis to obtain the content feedback activity of each video in the historical content set.

[0035] The picture clarity, picture stability and color reproduction of each video in the historically released content set are collected, and ratio analysis is performed with the preset reference picture clarity, picture stability and color reproduction. The results of each ratio analysis are accumulated to obtain the content viewing comfort of each video in the historically released content set. It should be noted that the picture clarity is the product of the frame rate and resolution of the video, the picture stability is the inverse of the average jitter amplitude during the simulated playback test of the video, and a preset constant is added to the denominator during the inversion process to prevent the denominator from being 0, and the color reproduction is the gamma value during the simulated playback test of the video.

[0036] It should be noted that the preset reference values ​​used in the above-mentioned analysis of the content information density, content feedback activity and content viewing comfort of each video in the historically released content set can be directly extracted from the cloud database, and each preset reference value is artificially developed and designed and implanted into the cloud database.

[0037] It should also be noted that the content quality coefficient of each video in the above-mentioned historically released content set is obtained by accumulating its content information density, content feedback activity and content viewing comfort, and the content creation quality coefficient of each video publishing author in the currently examined video set is obtained by averaging the content quality coefficients of each video in the historically released content set of each video publishing author in the currently examined video set.

[0038] The cumulative value of the publishing author's basic reputation coefficient and the content creation quality coefficient is used as the publishing author's reliability, thereby obtaining the publishing author's reliability of each video in the current video set.

[0039] Specifically, the analysis of the reliability of the publishing source of each video in the current investigation video set also includes: configuring a copyright detection tool through a target video platform to detect whether each video in the current investigation video set has the originality of the publishing author, thereby screening each video in the current investigation video set into original videos and non-original videos, tracing the publishing source author and publishing source platform of each non-original video, setting a reliability correction factor according to the publishing source platform, and then evaluating the reliability of the publishing source author of each non-original video.

[0040] It should be noted that the specific process of setting the reliability correction factor according to the publishing source platform is as follows: if the publishing source platform of the non-original video is the target video platform, then its reliability correction factor is set to If the source platform of the non-original video is not the target video platform, then set its reliability correction factor to ,in , thereby obtaining the corresponding reliability correction factor of the publishing source platform of each non-original video.

[0041] Consistent with the analysis method of the basic credibility coefficient and content creation quality coefficient of the publishing author, the basic credibility coefficient and content creation quality coefficient of the publishing source author of each non-original video are analyzed, and the cumulative value of the basic credibility coefficient and content creation quality coefficient of the publishing source author of each non-original video is further multiplied with the corresponding reliability correction factor of the publishing source platform, thereby obtaining the reliability of the publishing source author of each non-original video.

[0042] The reliability of the publishing author of each original video in the current video set is taken as its publishing source reliability, and the publishing author reliability and publishing source author reliability of each non-original video in the current video set are respectively added up by multiplying their corresponding preset weights to obtain the publishing source reliability of each non-original video in the current video set.

[0043] In summary, the reliability of the publishing source of each video in the current inspection video set is obtained.

[0044] Specifically, the analysis of the credibility of the published content of each video in the current video set includes: extracting keywords from the audio content of each video in the current video set, comparing the extracted keywords with the sensitive words in the target video platform standard sensitive word set stored in the cloud database, and counting the number of sensitive keywords of each video in the current video set.

[0045] Image elements are captured from the screen content of each video in the current video set under investigation, and each captured image element is compared with each sensitive image element in the target video platform standard sensitive image element set stored in the cloud database, so as to count the number of sensitive image elements of each video in the current video set under investigation.

[0046] According to the number of sensitive image elements in each video in the current video set and number of sensitive keywords , is the number of each video in the current video set. , according to the formula Analyze the content compliance coefficient of each video in the current video set, where They are respectively the total number of sensitive image elements in the target video platform's standardized sensitive image element set and the total number of sensitive words in the target video platform's standardized sensitive word set.

[0047] Similarly, the content information density, content feedback activity and content viewing comfort of each video in the current inspection video set are analyzed to analyze the content quality coefficient of each video in the current inspection video set.

[0048] The content quality coefficient and the content compliance coefficient are accumulated to obtain the credibility of the published content of each video in the current video set.

[0049] Specifically, the analysis of the positive trend of release and dissemination of each video in the current inspection video set includes: collecting the comment content in the comment area of ​​each video in the current inspection video set and evaluating its positive trend status index according to the positive trend vocabulary set and non-positive trend vocabulary set corresponding to the user comment retrieval specification stored in the cloud database, obtaining the comprehensive positive trend status index of the comment feedback of each video in the current inspection video set by cumulative calculation, and further multiplying it with the click-dislike ratio of the corresponding video in the current inspection video set as of the current time, to obtain the positive trend of release and dissemination of each video in the current inspection video set.

[0050] It should be noted that the positive trend state index of the above comment content takes the value of 1, -1 or 0, where 1 indicates that the comment content is in a positive trend state, -1 indicates that the comment content is in a non-positive trend state, and 0 indicates that the comment content is in a neutral trend state. The value of the positive trend state index depends on the comparison relationship between the number of positive trending words and the number of non-positive trending words in the comment content. If the number of positive trending words in the comment content is greater than the number of non-positive trending words, the positive trend state index of the output comment content is 1; if the number of positive trending words in the comment content is less than the number of non-positive trending words, the positive trend state index of the output comment content is -1; if the comment content does not identify positive trending words and non-positive trending words or the number of identified positive trending words is equal to the number of non-positive trending words, the positive trend state index of the output comment content is 0.

[0051] Specifically, the video browsing ranking of the currently examined video set is intervened once, including: accumulating the product of the publishing source reliability, publishing content credibility and publishing dissemination positivity of each video in the currently examined video set with their corresponding preset weights, to obtain the comprehensive information credibility of each video in the currently examined video set.

[0052] The videos in the current inspection video set whose comprehensive information credibility is less than the preset threshold of the comprehensive information credibility of the target video platform are eliminated, and the remaining videos are sorted in descending order according to the comprehensive information credibility, so as to obtain the video browsing ranking after one intervention in the current inspection video set.

[0053] The embodiment of the present invention analyzes the reliability of the publishing source, the credibility of the publishing content and the positive trend of publishing and dissemination of each video in the current video set, and conducts an intervention operation on the video browsing order of the current video set according to the principle of information credibility priority. This not only ensures that high-quality content obtains sufficient exposure opportunities and improves users' trust and satisfaction with the platform, but also effectively identifies and intercepts non-high-quality content, thereby improving the overall content quality of the platform's video push to users.

[0054] The video browsing ranking secondary intervention module is used to perform secondary intervention on the video browsing ranking of the current inspection video set according to the information diversity optimization principle.

[0055] Specifically, the secondary intervention on the video browsing order of the current inspection video set includes: determining the video grouping span corresponding to the single browsing cycle of the current inspection video set by the designated user, thereby dividing the current inspection video set after the one intervention into various inspection video groups.

[0056] Retrieve the target video platform's browsing preference weights for various video themes and video lengths obtained by analyzing the designated user's historical browsing behavior, and then obtain the designated user's browsing preference weights for various existing video themes and video lengths for the current inspection video set, and multiply them by the designated user's corresponding video grouping span for a single browsing cycle of the current inspection video set, to obtain the number of planned videos for each existing video theme and each length for a single inspection video group, and use them as the video theme planning requirements and video length planning requirements for the inspection video group, thereby performing secondary intervention on the video browsing order of the current inspection video set.

[0057] Specifically, the determination of the video group span corresponding to a single browsing period of the current video set by the designated user includes: collecting the number of video browsings of the designated user on each day within a historical preset period corresponding to the same time period in the future preset time period; , It is the number of each day in the historical preset period. , according to the formula Get the number of reference video views corresponding to the preset time period of the specified user in the future, where is the number of days in the historical preset period. The ratio of the number of reference video views corresponding to the future preset time period of the specified user to the preset planned grouping number is used as the video grouping span corresponding to the single browsing period of the specified user for the current inspection video set.

[0058] The embodiment of the present invention is based on the principle of information diversity optimization, combines the two dimensions of video theme type and video browsing length, and performs a secondary intervention operation on the video browsing ranking of the currently examined video set, which not only reduces invalid browsing caused by users' lack of interest or repeated and similar content, and improves the user's overall browsing experience, but also pays attention to the comprehensiveness of the content, avoids users from being trapped in information cocoons, and greatly enhances users' stickiness and loyalty to the platform.

[0059] The intervention effect evaluation module is used to use the video browsing ranking after the secondary intervention of the current inspection video set as the actual video playback ranking, collect the actual browsing behavior data of the specified user within a preset time period in the future of the day, evaluate the intervention effect of the video browsing ranking of the current inspection video set and provide feedback.

[0060] Specifically, the actual browsing behavior data includes the browsing time and interactive activity of each video in the currently examined video set.

[0061] Specifically, the intervention effect of evaluating the video browsing ranking of the current video set under investigation includes: comparing the actual browsing behavior data of the specified user in the future preset time period of the day with the actual browsing behavior data of the same time period corresponding to the future preset time period on each day of the historical preset period, and obtaining the year-on-year growth rate of the number of video browsing of the specified user in the future preset time period of the day , the year-on-year growth rate of interactive activity for various video topics in the current video collection And the year-on-year growth rate of viewing time of various video lengths , is the number of each video topic in the current video set. , is the number of the length of each type of video in the current video set. , according to the formula Get the intervention effect evaluation index of the video browsing ranking of the current video set, where The first Video-like topic browsing preference weight, The first Browsing preference weight of video length.

[0062] The cloud database is used to store the target video platform's standardized sensitive vocabulary set and sensitive image element set, and store the positive trending vocabulary set and the non-positive trending vocabulary set corresponding to the user comment retrieval standard.

[0063] It should be noted that the data sources of the target video platform standard sensitive vocabulary set and sensitive image element set, and the positive trend vocabulary set and non-positive trend vocabulary set corresponding to the user comment retrieval specification stored in the above-mentioned cloud database mainly rely on the manual preparation and uploading by the target platform managers.

[0064] The embodiments of the present invention help to more objectively and accurately evaluate the actual effects of video browsing ranking intervention measures by collecting actual browsing behavior data of designated users on the target video platform, and provide a scientific basis for the optimization and adjustment of subsequent platform push video browsing ranking intervention.

[0065] The above contents are merely examples and explanations of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, they shall all fall within the protection scope of the present invention.

Claims

1. An intelligent management system for information dissemination on a social platform based on human-machine symbiosis, characterized in that: The system includes: The module for extracting the video set to be pushed is used to extract the video set to be pushed by the target video platform for the specified user within a preset time period in the future on the same day, and record it as the current inspection video set; The video browsing ranking one-time intervention module is used to analyze the publishing source reliability, publishing content credibility and publishing dissemination positive trend of each video in the current inspection video set according to the principle of information credibility priority, so as to intervene in the video browsing ranking of the current inspection video set; The video browsing ranking secondary intervention module is used to perform secondary intervention on the video browsing ranking of the current inspection video set according to the information diversity optimization principle; The intervention effect evaluation module is used to use the video browsing ranking after the secondary intervention of the current inspection video set as the actual video playback ranking, collect the actual browsing behavior data of the specified user in the future preset time period of the day, evaluate the intervention effect of the video browsing ranking of the current inspection video set and provide feedback; The cloud database is used to store the target video platform's standardized sensitive vocabulary set and sensitive image element set, and to store the positive trending vocabulary set and non-positive trending vocabulary set corresponding to the user comment retrieval specification.

2. According to claim 1, a social platform information dissemination intelligent management system based on human-machine symbiosis is characterized by: The analysis of the reliability of the publishing source of each video in the current video set includes: collecting the publishing author information of each video in the current video set, including the official certification index of the target video platform, the registration years, the community participation contribution, the number of violations and the severity of each violation, the number of fans, the fan activity and the fan retention rate, and analyzing the influence, recognition and risk of the publishing author information respectively, so as to evaluate the basic credibility coefficient of each video publishing author in the current video set; Retrieve the historical published content set of the publisher of each video in the currently examined video set, analyze the content information density, content feedback activity and content viewing comfort of each video in the historical published content set, and analyze the content quality coefficient of each video in the historical published content set, so as to evaluate the content creation quality coefficient of each video publisher in the currently examined video set; The cumulative value of the publishing author's basic reputation coefficient and the content creation quality coefficient is used as the publishing author's reliability, thereby obtaining the publishing author's reliability of each video in the current video set.

3. According to claim 2, a social platform information dissemination intelligent management system based on human-machine symbiosis is characterized by: The analysis of the reliability of the publishing source of each video in the current video set includes: configuring a copyright detection tool on the target video platform to detect whether each video in the current video set has the originality of the publishing author, thereby screening each video in the current video set into original videos and non-original videos, tracing the publishing source author and publishing source platform of each non-original video, setting a reliability correction factor according to the publishing source platform, and then evaluating the reliability of the publishing source author of each non-original video; The reliability of the author of each original video in the current video set is used as its release source reliability, and the reliability of the author of each non-original video in the current video set is accumulated by multiplying the product of the author reliability and the release source author reliability of each non-original video in the current video set with their corresponding preset weights to obtain the release source reliability of each non-original video in the current video set; In summary, the reliability of the publishing source of each video in the current inspection video set is obtained.

4. According to claim 1, a social platform information dissemination intelligent management system based on human-machine symbiosis is characterized by: The analyzing the credibility of the published content of each video in the current video set includes: extracting keywords from the audio content of each video in the current video set, comparing the extracted keywords with the sensitive words in the target video platform standard sensitive word set stored in the cloud database, and counting the number of sensitive keywords of each video in the current video set; Image elements are intercepted from the screen content of each video in the current video set, and each intercepted image element is compared with each sensitive image element in the target video platform standard sensitive image element set stored in the cloud database, so as to count the number of sensitive image elements of each video in the current video set; Analyze the content compliance coefficient of each video in the current video set based on the number of sensitive image elements and sensitive keywords; Similarly, the content information density, content feedback activity and content viewing comfort of each video in the current video set are analyzed to analyze the content quality coefficient of each video in the current video set; The content quality coefficient and the content compliance coefficient are accumulated to obtain the credibility of the published content of each video in the current video set.

5. According to claim 1, a social platform information dissemination intelligent management system based on human-machine symbiosis is characterized by: The analysis of the positive trend of release and dissemination of each video in the current inspection video set includes: collecting the comment content in the comment area of ​​each video in the current inspection video set and evaluating its positive trend state index according to the positive trend vocabulary set and the non-positive trend vocabulary set corresponding to the user comment retrieval specification stored in the cloud database, obtaining the comprehensive positive trend state index of the comment feedback of each video in the current inspection video set by cumulative calculation, and further multiplying it with the click-dislike ratio of the corresponding video in the current inspection video set as of the current time, to obtain the positive trend of release and dissemination of each video in the current inspection video set.

6. The social platform information dissemination intelligent management system based on human-machine symbiosis according to claim 1 is characterized by: The intervention in the video browsing ranking of the current inspection video set includes: accumulating the product of the release source reliability, release content credibility and release dissemination positive trend of each video in the current inspection video set and its corresponding preset weight respectively, to obtain the comprehensive information credibility of each video in the current inspection video set; The videos in the current inspection video set whose comprehensive information credibility is less than the preset threshold of the comprehensive information credibility of the target video platform are eliminated, and the remaining videos are sorted in descending order according to the comprehensive information credibility, so as to obtain the video browsing ranking after one intervention in the current inspection video set.

7. According to claim 1, a social platform information dissemination intelligent management system based on human-machine symbiosis is characterized by: The second intervention on the video browsing order of the current inspection video set includes: determining the video grouping span corresponding to the single browsing cycle of the current inspection video set by the designated user, thereby dividing the current inspection video set after the one intervention into various inspection video groups; Retrieve the target video platform's browsing preference weights for various video themes and video lengths obtained by analyzing the designated user's historical browsing behavior, and then obtain the designated user's browsing preference weights for various existing video themes and video lengths for the current inspection video set, and multiply them by the designated user's corresponding video grouping span for a single browsing cycle of the current inspection video set, to obtain the number of planned videos for each existing video theme and each length for a single inspection video group, and use them as the video theme planning requirements and video length planning requirements for the inspection video group, thereby performing secondary intervention on the video browsing order of the current inspection video set.

8. The social platform information dissemination intelligent management system based on human-machine symbiosis according to claim 7 is characterized by: Determining the video group span corresponding to a single browsing period of the current video set for the designated user includes: collecting the number of video browsings of the designated user on each day of the historical preset period corresponding to the same time period for the future preset time period , It is the number of each day in the historical preset period. , according to the formula Get the number of reference video views corresponding to the preset time period of the specified user in the future, where is the number of days in the historical preset period. The ratio of the number of reference video views corresponding to the future preset time period of the specified user to the preset planned grouping number is used as the video grouping span corresponding to the single browsing period of the specified user for the current inspection video set.

9. The social platform information dissemination intelligent management system based on human-machine symbiosis according to claim 8, characterized in that: The actual browsing behavior data includes the browsing time and interactive activity of each video in the currently examined video set.

10. The social platform information dissemination intelligent management system based on human-machine symbiosis according to claim 9, characterized in that: The intervention effect of evaluating the video browsing ranking of the current video set under investigation includes: comparing the actual browsing behavior data of the specified user in the future preset time period of the day with the actual browsing behavior data of the same time period corresponding to the future preset time period on each day of the historical preset period, and obtaining the year-on-year growth rate of the number of video browsing of the specified user in the future preset time period of the day. , the year-on-year growth rate of interactive activity for various video topics in the current video collection And the year-on-year growth rate of viewing time of various video lengths , is the number of each video topic in the current video set. , is the number of the length of each type of video in the current video set. , according to the formula Get the intervention effect evaluation index of the video browsing ranking of the current video set, where The first Video-like topic browsing preference weight, The first Browsing preference weight of video length.

Images