An intelligent cultural science popularization platform management system and management method

Abstract

The present invention discloses an intelligent cultural science popularization platform management system and management method, which relates to the field of platform management technology. The system first collects diversified cultural science popularization resources from the Internet through a cultural data capture module to establish a cultural database. The combination of the user behavior acquisition module and the first analysis module can deeply analyze the personalized characteristics and behavior patterns of users. The user's knowledge acceptance Jsd is calculated and evaluated, which enables the platform to dynamically adjust the priority of pushed content according to the user's interests and behaviors. The second analysis module calculates the social activity Shy by real-time monitoring of the user's social interaction data, and corrects the push strategy in combination with the push level. The relevance of the pushed content and the user's participation are ensured, thereby improving the user's acceptance and satisfaction with the content. By analyzing the push content willingness Yyd, the push frequency and content type are adjusted, which improves the platform's responsiveness and user experience.

Description

Technical Field

[0001] The present invention relates to the technical field of platform management, and in particular to an intelligent cultural science popularization platform management system and a management method. Background Art

[0002] The main function of cultural science platforms is to enhance the public's knowledge and interest in various cultural knowledge through the dissemination of cultural content in various forms. These platforms usually provide rich cultural resources, including knowledge in the fields of history, literature, art, ethnicity, etc., and display them in various forms such as text, images, audio, and video. Through these platforms, users can easily obtain cultural information and understand the background and value of different cultures, thereby promoting the improvement of cultural literacy and cultural inheritance.

[0003] However, the existing traditional culture popular science platform has some shortcomings in practice, which affect the effectiveness of the platform and user experience. Specifically, the existing platform's analysis of user behavior is mainly focused on basic data such as page views and clicks, and fails to deeply explore users' personalized needs and behavior patterns. This limits the platform's optimization in content push and user interaction, making it difficult to provide accurate content and services. The feedback mechanism of the existing platform usually cannot capture and respond to user feedback on pushed content in real time. This makes the platform slow to respond in adjusting push strategies and optimizing content, and it is difficult to quickly adapt to users' changing needs. Summary of the invention

[0004] In view of the deficiencies in the prior art, the present invention provides an intelligent cultural science popularization platform management system and management method to solve the problems mentioned in the background technology.

[0005] To achieve the above objectives, the present invention is implemented through the following technical solutions: an intelligent cultural science platform management system, including a cultural data capture module, a processing module, a user behavior collection module, a first analysis module, a first push module and a second analysis module;

[0006] The cultural data crawling module is used to collect several types of cultural popular science content from public network resources on the Internet through Web crawler tools, and to establish a cultural database. After pre-processing the cultural database through the processing module, it is integrated and transmitted on the popular science platform;

[0007] The user behavior collection module is used to collect the content and behavior data browsed by users on the popular science platform and establish a first behavior data set;

[0008] The first analysis module is used to extract the personalized characteristics, behavioral pattern characteristics and psychological characteristics of the user in the first behavioral data set, and perform in-depth calculations through the K-means clustering algorithm and the Apriori algorithm to obtain the user's cultural classification jump count CCTi , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ; and classify the user culture into jump times CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i Associated to calculate the user knowledge acceptance Jsd;

[0009] A first push module is used to preset a first threshold value X1 and a second threshold value X2, and compare the first threshold value X1 and the second threshold value X2 with the user knowledge acceptance degree Jsd respectively to obtain a corresponding push level result, and generate a first push instruction according to the corresponding push level result;

[0010] The second analysis module is used to collect and obtain user social interaction data in real time and extract social interaction features to calculate the user's social activity Shy, and compare and analyze the social activity Shy with the third threshold value X3 to obtain the user activity evaluation result, and associate the user activity evaluation result with the corresponding push level to correct the first push instruction, obtain the second push instruction and execute it.

[0011] Preferably, the cultural data capture module includes a multi-source data capture unit and a multi-type content classification unit;

[0012] The multi-source data capture unit is used to collect several types of cultural popular science content from public network resources on the Internet through Web crawler tools, and establish a cultural database, and perform semantic analysis on the captured cultural data, identify the subject and keywords of the content through natural language processing (NLP) technology, and intelligently associate the relevant cultural data, and use text classification algorithms to classify the cultural database by subject, and automatically generate tags according to the content to obtain the subject classification tags;

[0013] The multi-type content classification unit is used to classify the cultural database to obtain several groups of content types, including text, image, audio, video, AR / VR content, documentary materials, historical archives and oral history;

[0014] The processing module is used to preprocess the cultural database, which includes data cleaning, data standardization, text processing and data conversion, and to associate different types of data, including associating text content with corresponding pictures and audio, and after aggregation processing, transmit it to the popular science platform for display.

[0015] Preferably, the first analysis module includes a first extraction unit and a user behavior analysis unit;

[0016] The first extraction unit is used to extract the browsing behavior, clicking behavior, page jump behavior and scrolling behavior characteristics of the user on the cultural science platform from the behavior data set, and establish user personalized characteristics, behavior pattern characteristics and psychological characteristics;

[0017] The user behavior analysis unit is used to analyze and calculate based on the user's personalized characteristics, behavior pattern characteristics and psychological characteristics to obtain: the number of user cultural classification jumps CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ;

[0018] User culture classification jump times CCT i Calculated by the following formula:

[0019]

[0020] In the formula, C j represents the number of times a user jumps from one cultural category to another in the i-th time period, and m is the number of cultural categories visited by the user in the time period;

[0021] User Comprehensive Rolling Index CSI i Calculated by the following formula;

[0022] CSI i =S i *D i ;

[0023] In the formula, S i is the page scrolling speed of the user during browsing in the i-th time period, D i is the page scrolling depth of the user's browsing process during the i-th time period;

[0024] User culture classification residence time CST i Calculated by the following formula:

[0025]

[0026] Where, T k represents the time the user stays on the kth cultural category page, and m is the number of cultural categories visited by the user during this time period;

[0027] And the user culture classification jump number CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i The user knowledge acceptance Jsd is calculated by the following correlation formula:

[0028]

[0029] In the formula, n represents the number of time periods into which the user behavior data is divided, and i represents the data of the i-th time period; B i Indicates the frequency of the user's retrospective behavior in the i-th time period.

[0030] Preferably, the first pushing module includes a first evaluation unit and a first pushing unit;

[0031] The first evaluation unit is used to preset a first threshold value X1 and a second threshold value X2, and compare the first threshold value X1 and the second threshold value X2 with the user knowledge acceptance Jsd respectively to obtain a corresponding push level result, including:

[0032] When the user knowledge acceptance Jsd is less than the first threshold X1, the first acceptance level is obtained, indicating that the user's interest in the current content is less than 30%;

[0033] When the first threshold X1≤user knowledge acceptance Jsd≤the second threshold X2, the second acceptance level is obtained, indicating that the user's interest in the current content is between 31% and 69%;

[0034] When the user knowledge acceptance Jsd> the second threshold X2, the third acceptance level is obtained, indicating that the user's interest in the current content is between 70% and 100%;

[0035] The first push unit is used to generate a first push instruction according to the corresponding push level result, including:

[0036] Generate a first low-priority content push instruction based on the first acceptance level, including pushing a short prompt message, title guide or summary content, and set the push frequency to twice a week through a single message channel within the platform;

[0037] Generate a second intermediate priority content push instruction based on the second acceptance level, including pushing short articles, picture collections or FAQs, and set the push frequency to 4 times a week, and push them simultaneously through platform messages and email push;

[0038] Generate third-highest priority content push instructions based on the third acceptance level, including pushing in-depth articles, videos or expert recommendation materials, and set the push frequency to 6 times a week; and push them simultaneously through platform messages, emails and APP notifications.

[0039] Preferably, the second analysis module includes a social activity data collection unit and a social interaction analysis unit;

[0040] The social activity data collection unit is used to obtain user social interaction data through real-time collection; embed tracking scripts in the platform page to monitor user clicks and mouse hover interaction behaviors, capture real-time interaction data, and obtain user social interaction data; user social interaction data includes likes, comments, shares, discussion group participation, and friend interaction information;

[0041] The social interaction analysis unit is used to extract social interaction features from user social interaction data and perform in-depth calculations to obtain the user interaction time fluctuation index ITVI. i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i ;

[0042] User interaction time fluctuation index ITVI i Acquisition includes the following steps:

[0043] Collect the user's interaction time data in the i-th time period and mark it as: i1 , L i2 ,...,L ig}where g is the number of interactive data points in the time period;

[0044] Calculate the average user interaction time during this time period

[0045]

[0046] Where, L ij is the time of the jth interaction in the i-th time period;

[0047] Calculate the variance of user interaction time during this time period

[0048]

[0049] In the formula, the variance It represents the average of the sum of squares of the deviations of the data points from the mean;

[0050] ITVI, the user interaction time fluctuation index i Calculated by the following formula:

[0051]

[0052] User Interaction Delay Time Index IDTI i Calculated by the following formula:

[0053]

[0054] In the formula, ΔYc ij is the latency of the jth interaction in the i-th time period, the time from when the user initiates the interaction to when the response is received; g is the number of interaction data points in the time period;

[0055] User social influence index Sll i Calculated by the following formula:

[0056]

[0057] In the formula, E i is the number of times the user interaction content is cited or forwarded by other users in the i-th time period; R i is the total number of user interactions in the i-th time period;

[0058] User Interaction Diversity Index IDI i Calculated by the following formula:

[0059]

[0060] In the formula, F ij is the frequency of the jth social interaction type in the i-th time period, and the social interaction types include comments, likes, and shares; T ij is the total number of interactions of the jth type in the ith time period; e is the total number of interaction types;

[0061] User interaction time distribution ITDI i Calculated by the following formula:

[0062]

[0063] In the formula, Entropy(H i ) is the entropy value of the user interaction time distribution in the i-th time period, H max is the maximum entropy value.

[0064] Preferably, the second analysis module further includes a user social activity calculation unit, a second evaluation unit and a first correction unit;

[0065] The user social activity calculation unit is used to calculate the user interaction time fluctuation index ITVI i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i After dimensionless processing, the user's social activity Shy is calculated using the following associated formula:

[0066]

[0067] Where a1, a2, a3, a4 and a5 represent the user interaction time fluctuation index ITVI i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i The weight value of b1 represents the first correction constant value;

[0068] The second evaluation unit is used to preset a third threshold value X3, and compare the third threshold value X3 with the user's social activity Shy to obtain a user activity evaluation result, including:

[0069] When the user's social activity Shy is less than the third threshold value X3, it means that the user's social activity is unqualified, and a first unqualified label is generated, indicating that the user is in a passive state in social interaction;

[0070] When the user's social activity Shy ≥ the third threshold value X3, it means that the user's social activity is qualified, and a first qualified label is generated, indicating that the user is in an active state in social interaction;

[0071] The first correction unit is used to associate the user activity evaluation result with the corresponding push level to correct the first push instruction and obtain the second push instruction, including:

[0072] If a user who receives the first low-priority content push instruction generates the first unqualified label, the push frequency setting is changed to once a week; if a user who receives the second medium-priority content push instruction generates the first unqualified label, the push frequency setting is changed to three times a week; if a user who receives the third high-priority content push instruction generates the first unqualified label, the push frequency setting is changed to five times a week;

[0073] For users who receive the first low-priority content push instruction and generate the first qualified tag, the push plan will be short content and guiding titles, and the push frequency will be maintained at 2 times a week, which will be pushed through a single channel of platform messages; for users who receive the second intermediate-priority content push instruction and generate the first qualified tag, the push plan will be a combination of pictures and texts and interactive questions and answers, and the push frequency will be maintained at 4 times a week, which will be pushed simultaneously through platform messages and emails; for users who receive the third-highest-priority content push instruction and generate the first qualified tag, the settings will be changed to in-depth content and discussion group invitations, and the push frequency will be maintained at 6 times a week, which will be pushed simultaneously through platform messages, emails and APP notifications.

[0074] Preferably, it also includes a push feedback detection module, which includes a push feedback information collection unit and a push willingness calculation unit;

[0075] The push feedback information collection unit is used to collect various behavior data related to the user and the push content, establish a push feedback data set, and the push willingness calculation unit extracts the interaction characteristics between the user and the push content in the push feedback data set to calculate the user push acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i ; and push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The related calculation obtains the push content willingness Yyd;

[0076] User push acceptance rate PAP i Calculated by the following formula:

[0077]

[0078] In the formula, Js i Ts represents the number of pushed content that users accept and click in the i-th time period. i Indicates the total number of push contents received by the user in the i-th time period;

[0079] Interactive Feedback Rate (IFR) i Calculated by the following formula:

[0080]

[0081] Where Hdcs i Indicates the total number of user interactions with the pushed content in the i-th time period, including the total number of likes, comments, and shares;

[0082] RFT, the user's feedback delay index for pushed content i Calculated by the following formula:

[0083]

[0084] In the formula, Fksj ij Indicates the feedback time after receiving the jth push in the i-th time period;

[0085] Push content relevance index PCR iCalculated by the following formula:

[0086]

[0087] In the formula, XG ij Indicates the relevance score between the jth pushed content and the user's interests in the i-th time period;

[0088] And push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i After dimensionless processing, the push content willingness Yyd is calculated through the following related formula:

[0089]

[0090] Where a6, a7, a8 and a9 represent the user push acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The weight value of b2 represents the second correction constant value.

[0091] Preferably, the push feedback detection module includes a third evaluation unit and a second correction unit; the third evaluation unit is used to preset a fourth threshold value X4, and compare and analyze the fourth threshold value X4 with the push content willingness Yyd to obtain a push feedback evaluation result, including:

[0092] When the push content willingness Yyd is less than the fourth threshold value X4, it indicates that the user is dissatisfied with the push feedback, and a first feedback abnormality label is generated;

[0093] When the push content willingness Yyd ≥ the fourth threshold X4, it means that the user is satisfied with the push feedback, and a first feedback normal label is generated;

[0094] The second correction unit is used to generate a corresponding adjustment strategy according to the push feedback evaluation result, including:

[0095] If the first feedback normal tag is generated, the specific content in the second push instruction is maintained;

[0096] If the first feedback abnormal tag is generated, the second push instruction is adjusted, including:

[0097] The first low-priority content push instruction adjusts the push frequency to once a week. The push content is still short content and guiding headlines, and is pushed through a single channel of messages within the platform;

[0098] The second intermediate priority content push instruction adjusts the push frequency to 3 times a week. The push content is a combination of pictures and texts and interactive questions and answers, and is pushed simultaneously through platform messages and email push;

[0099] The third high-level priority content push instruction adjusts the push frequency to 5 times a week. The push content includes in-depth content and discussion group invitations, and is pushed simultaneously through platform messages, emails and APP notifications.

[0100] A method for managing an intelligent cultural science popularization platform comprises the following steps:

[0101] S1. Use web crawler tools to collect several types of cultural and popular science content from public network resources on the Internet, and establish a cultural database. After pre-processing the cultural database through a processing module, integrate and transmit it on the popular science platform;

[0102] S2, collecting the content and behavior data browsed by users on the popular science platform, and establishing a first behavior data set;

[0103] S3: Extract the personalized features, behavioral pattern features, and psychological features of the user in the first behavioral data set, and perform deep calculations using the K-means clustering algorithm and the Apriori algorithm to obtain the number of cultural classification jumps CCT of the user i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ; and classify the user culture into jump times CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i The associated calculation obtains the user's knowledge acceptance Jsd;

[0104] And evaluate the user knowledge acceptance Jsd to obtain the corresponding push level result, and generate the first push instruction according to the corresponding push level result;

[0105] S4, collecting and acquiring user social interaction data in real time and extracting social interaction features to calculate and acquire user social activity Shy, and evaluating user social activity Shy to acquire user activity evaluation results, and associating user activity evaluation results with corresponding push levels to correct the first push instruction, and acquire and execute the second push instruction;

[0106] S5. Collect various behavior data related to users and pushed content, establish a push feedback data set, and extract the interaction features between users and pushed content in the push feedback data set to calculate the user push acceptance rate PAP i , user interaction feedback rate IFR i, User feedback delay index RFT for pushed content i PCR index of push content relevance i ; and push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The push content willingness degree Yyd is obtained by associated calculation; and the push content willingness degree Yyd is evaluated to generate a push feedback evaluation result, and a corresponding adjustment strategy is generated based on the push feedback evaluation result.

[0107] The present invention provides an intelligent cultural science popularization platform management system and management method. It has the following beneficial effects:

[0108] The system extracts the personalized features, behavioral pattern features and psychological features of the user in the first behavioral data set, analyzes and obtains the user's knowledge acceptance Jsd, and can accurately match the user's interests and needs through personalized analysis and intelligent recommendation algorithms, thereby improving the accuracy of content recommendations and enhancing user experience. The system collects and analyzes the user's social interaction data in real time, extracts social interaction features, calculates and evaluates the user's social activity Shy, obtains the user activity evaluation result, and associates the user activity evaluation result with the corresponding push level to correct the first push instruction, obtain the second push instruction and execute it. The platform can capture user feedback information in a timely manner and quickly adjust the push strategy, thereby improving the response speed and content relevance. By deeply analyzing the user's behavioral data and social interaction features, the platform can better understand user needs, optimize content and interaction strategies, and increase user participation and satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0109] Figure 1 This is a flowchart diagram of an intelligent cultural science platform management system of the present invention;

[0110] Figure 2 This is a schematic diagram of the steps of a method for managing an intelligent cultural science popularization platform of the present invention. DETAILED DESCRIPTION

[0111] 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.

[0112] Example 1

[0113] See also Figure 1 , the present invention provides an intelligent cultural science popularization platform management system, including a cultural data capture module, a processing module, a user behavior collection module, a first analysis module, a first push module and a second analysis module;

[0114] The cultural data crawling module is used to collect several types of cultural popular science content from public network resources on the Internet through Web crawler tools, and to establish a cultural database. After pre-processing the cultural database through the processing module, it is integrated and transmitted on the popular science platform;

[0115] The user behavior collection module is used to collect the content and behavior data browsed by users on the popular science platform and establish a first behavior data set;

[0116] The first analysis module is used to extract the personalized characteristics, behavioral pattern characteristics and psychological characteristics of the user in the first behavioral data set, and perform in-depth calculations through the K-means clustering algorithm and the Apriori algorithm to obtain the user's cultural classification jump count CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ; and classify the user culture into jump times CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i Associated to calculate the user knowledge acceptance Jsd;

[0117] A first push module is used to preset a first threshold value X1 and a second threshold value X2, and compare the first threshold value X1 and the second threshold value X2 with the user knowledge acceptance degree Jsd respectively to obtain a corresponding push level result, and generate a first push instruction according to the corresponding push level result;

[0118] The second analysis module is used to collect and obtain user social interaction data in real time and extract social interaction features to calculate the user's social activity Shy, and compare and analyze the social activity Shy with the third threshold value X3 to obtain the user activity evaluation result, and associate the user activity evaluation result with the corresponding push level to correct the first push instruction, obtain the second push instruction and execute it.

[0119] In this embodiment, the system extracts the personalized characteristics, behavioral pattern characteristics and psychological characteristics of the user in the first behavioral data set, analyzes and obtains the user's knowledge acceptance Jsd, and can accurately match the user's interests and needs through personalized analysis and intelligent recommendation algorithms, thereby improving the accuracy of content recommendation and enhancing user experience. The user's social interaction data is collected and analyzed in real time, and social interaction characteristics are extracted to calculate and evaluate the user's social activity Shy to obtain the user's activity evaluation result, and the user activity evaluation result is associated with the corresponding push level to correct the first push instruction, obtain the second push instruction and execute it. The platform can capture user feedback information in a timely manner and quickly adjust the push strategy, thereby improving the response speed and content relevance. By deeply analyzing the user's behavioral data and social interaction characteristics, the platform can better understand user needs, optimize content and interaction strategies, and increase user participation and satisfaction.

[0120] Example 2, please refer to Figure 1 ,The cultural data crawling module includes a multi-source data crawling unit and a multi-type content classification unit;

[0121] The multi-source data capture unit is used to collect several types of cultural popular science content from the Internet through the Web crawler tool, and establish a cultural database, and perform semantic analysis on the captured cultural data, identify the theme and keywords of the content through natural language processing (NLP) technology, and intelligently associate the relevant cultural data, and use the text classification algorithm to subject-classify the cultural database, and automatically generate labels according to the content to obtain subject classification labels; the subject classification labels include historical culture, literature and art, national culture, natural heritage, scientific and technological inventions, and local culture. The local culture category includes cultural and historical background data, cultural relics and architectural data, tourist customs data, and scenic area geographical distribution data, and uses GIS technology to collect and analyze the geographical distribution, tourist flow and popular attractions of the scenic area, and combines cultural heritage information to form a cultural map;

[0122] The multi-type content classification unit is used to classify the cultural database and obtain several groups of content types, including text, images, audio, video, AR / VR content, documentary materials, historical archives and oral history; to provide users with diversified cultural experiences and improve the interactivity and fun of the content.

[0123] The processing module is used to preprocess the cultural database, which includes data cleaning, data standardization, text processing and data conversion, and to associate different types of data, including associating text content with corresponding pictures and audio, and after aggregation processing, transmit it to the popular science platform for display.

[0124] In this embodiment, through the multi-source data capture unit and intelligent association, the system can extract and integrate rich cultural data from a variety of network resources to form a rich cultural database. The comprehensiveness and accuracy of cultural data are improved, and users can obtain more cultural information. Cultural content is intelligently classified using NLP technology and text classification algorithms, and detailed topic tags are generated. The content classification on the platform is made clearer and more accurate, and users can more easily find content of interest. Local cultural data is combined with GIS technology to form a cultural map, providing information on the geographical distribution of scenic spots, tourist traffic and popular attractions. Help users better understand and plan cultural tourism and enhance the travel experience. Through data association and a variety of content display forms, improve user participation and interest in cultural content. Enable users to gain in-depth understanding of cultural knowledge through different forms, and enhance the interactivity and fun of learning.

[0125] Example 3, please refer to Figure 1 , the first analysis module includes a first extraction unit and a user behavior analysis unit;

[0126] The first extraction unit is used to extract the browsing behavior, clicking behavior, page jump behavior and scrolling behavior characteristics of the user on the cultural science platform from the behavior data set, and establish user personalized characteristics, behavior pattern characteristics and psychological characteristics;

[0127] The user behavior analysis unit is used to analyze and calculate based on the user's personalized characteristics, behavior pattern characteristics and psychological characteristics to obtain: the number of user cultural classification jumps CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ;

[0128] User culture classification jump times CCT i Calculated by the following formula:

[0129]

[0130] In the formula, C j represents the number of times a user jumps from one cultural category to another in the i-th time period, and m is the number of cultural categories visited by the user in the time period;

[0131] User Comprehensive Rolling Index CSI i Calculate it using the following formula;

[0132] CSI i =S i *D i ;

[0133] In the formula, S iis the page scrolling speed of the user during browsing in the i-th time period, D i is the page scrolling depth of the user's browsing process during the i-th time period;

[0134] User culture classification residence time CST i Calculated by the following formula:

[0135]

[0136] Where, T k represents the time the user stays on the kth cultural category page, and m is the number of cultural categories visited by the user during this time period;

[0137] And the user culture classification jump number CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i The user knowledge acceptance Jsd is calculated by the following correlation formula:

[0138]

[0139] In the formula, n represents the number of time periods into which the user behavior data is divided, and i represents the data of the i-th time period; B i represents the frequency of the user's retrospective behavior in the i-th time period, CCT i It reflects the user's jumping behavior between different cultural categories and indicates the user's breadth of exploration in cultural categories: CSI i Reflects the user's level of interaction on the page; CST i It represents the time users spend in a certain cultural category, summarizes the degree of users' attention to a certain cultural category, balances the time users spend in various cultural categories, and avoids excessive impact of frequent jumps or backtracking in a short period of time on the final acceptance.

[0140] The first pushing module includes a first evaluation unit and a first pushing unit;

[0141] The first evaluation unit is used to preset a first threshold value X1 and a second threshold value X2, and compare the first threshold value X1 and the second threshold value X2 with the user knowledge acceptance Jsd respectively to obtain a corresponding push level result, including:

[0142] When the user knowledge acceptance Jsd is less than the first threshold X1, the first acceptance level is obtained, indicating that the user's interest in the current content is less than 30%;

[0143] When the first threshold X1≤user knowledge acceptance Jsd≤the second threshold X2, the second acceptance level is obtained, indicating that the user's interest in the current content is between 31% and 69%;

[0144] When the user knowledge acceptance Jsd> the second threshold X2, the third acceptance level is obtained, indicating that the user's interest in the current content is between 70% and 100%;

[0145] The first push unit is used to generate a first push instruction according to the corresponding push level result, including:

[0146] Generate a first low-priority content push instruction based on the first acceptance level, including pushing a short prompt message, title guide or summary content, and set the push frequency to twice a week through a single message channel within the platform;

[0147] Generate a second intermediate priority content push instruction based on the second acceptance level, including pushing short articles, picture collections or FAQs, and set the push frequency to 4 times a week, and push them simultaneously through platform messages and email push;

[0148] Generate third-highest priority content push instructions based on the third acceptance level, including pushing in-depth articles, videos or expert recommendation materials, and set the push frequency to 6 times a week; and push simultaneously through platform messages, emails and APP notifications.

[0149] In this embodiment, through the first extraction unit and the user behavior analysis unit, the system can analyze the user's browsing, clicking, jumping and scrolling behaviors in detail. Providing accurate user behavior data helps to understand the user's interests and behavior patterns. Based on the results of the user behavior analysis, the system can provide users with personalized content recommendations based on the user knowledge acceptance Jsd. The push strategy of the first evaluation unit and the first push unit can adjust the push content and frequency according to the user's acceptance level. The effect of content push is optimized to ensure that users can receive suitable content at different interest stages, avoiding push interference and information overload. Through reasonable push frequency and diversified push channels, different levels of content push are provided. The user experience is enhanced so that users can obtain content that better suits their needs when using the cultural science platform.

[0150] Example 4, please refer to Figure 1 ,The second analysis module includes a social activity data collection unit and a social interaction analysis unit;

[0151] The social activity data collection unit is used to obtain user social interaction data through real-time collection; embed tracking scripts in the platform page to monitor user clicks and mouse hover interaction behaviors, capture real-time interaction data, and obtain user social interaction data; user social interaction data includes likes, comments, shares, discussion group participation, and friend interaction information;

[0152] The social interaction analysis unit is used to extract social interaction features from user social interaction data and perform in-depth calculations to obtain the user interaction time fluctuation index ITVI. i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i ;

[0153] ITVI, the user interaction time fluctuation index i Acquisition includes the following steps:

[0154] Collect the user's interaction time data in the i-th time period and mark it as: i1 , L i2 ,...,L ig}where g is the number of interactive data points in the time period;

[0155] Calculate the average user interaction time during this time period

[0156]

[0157] Where, L ij is the time of the jth interaction in the i-th time period;

[0158] Calculate the variance of user interaction time during this time period

[0159]

[0160] In the formula, the variance It represents the average of the sum of squares of the deviations of the data points from the mean;

[0161] ITVI, the user interaction time fluctuation index i Calculated by the following formula:

[0162]

[0163] User Interaction Delay Time Index IDTI i Calculated by the following formula:

[0164]

[0165] In the formula, ΔYc ij is the latency of the jth interaction in the i-th time period, the time from when the user initiates the interaction to when the response is received; g is the number of interaction data points in the time period;

[0166] User social influence index Sll i Calculated by the following formula:

[0167]

[0168] In the formula, E i is the number of times the user interaction content is cited or forwarded by other users in the i-th time period; R i is the total number of user interactions in the i-th time period;

[0169] User Interaction Diversity Index IDI i Calculated by the following formula:

[0170]

[0171] In the formula, F ij is the frequency of the jth social interaction type in the i-th time period, and the social interaction types include comments, likes, and shares; T ij is the total number of interactions of the jth type in the ith time period; e is the total number of interaction types;

[0172] User interaction time distribution ITDI i Calculated by the following formula:

[0173]

[0174] In the formula, Entropy(H i ) is the entropy value of the user interaction time distribution in the i-th time period, H max is the maximum entropy value.

[0175] The second analysis module also includes a user social activity calculation unit, a second evaluation unit and a first correction unit;

[0176] The user social activity calculation unit is used to calculate the user interaction time fluctuation index ITVI i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i After dimensionless processing, the user's social activity Shy is calculated using the following associated formula:

[0177]

[0178] Where a1, a2, a3, a4 and a5 represent the user interaction time fluctuation index ITVI i , User Interaction Delay Time Index IDTI i, User Social Influence Index Sll i , User Interaction Diversity Index IDI i and User Interaction Time Distribution ITDI i The weight values of are adjusted and set by the user. 0 < a1 < 1, 0 < a2 < 1, 0 < a3 < 1, 0 < a4 < 1, 0 < a5 < 1, and a1 + a2 + a3 + a4 + a5 = 1. b1 represents the first corrected value;

[0179] The second evaluation unit is used to preset the third threshold X3, and compare the third threshold X3 with the user's social activity Shy to obtain the user activity evaluation result, including:

[0180] When the user's social activity Shy < the third threshold X3, it means that the user's social activity is unqualified, and a first unqualified label is generated, indicating that the user is passive in social interaction;

[0181] When the user's social activity Shy ≥ the third threshold X3, it means that the user's social activity is qualified, and a first qualified label is generated, indicating that the user is active in social interaction;

[0182] The first correction unit is used to associate the user activity evaluation result with the corresponding push level to correct the first push instruction and obtain the second push instruction, including:

[0183] For users who obtain the first low-priority content push instruction, if a first unqualified label is generated, the push frequency is changed to once a week; for users who obtain the second medium-priority content push instruction, if a first unqualified label is generated, the push frequency is changed to three times a week; for users who obtain the third high-priority content push instruction, if a first unqualified label is generated, the push frequency is changed to five times a week;

[0184] For users who obtain the first low-priority content push instruction, if a first qualified label is generated, the push plan is short content and a guiding title, and the push frequency is maintained at twice a week, and it is pushed through a single channel of in-platform messages; for users who obtain the second medium-priority content push instruction, if a first qualified label is generated, the push plan is a combination of pictures and texts and interactive Q&A, and the push frequency is maintained at four times a week, and it is pushed synchronously through platform messages and emails; for users who obtain the third high-priority content push instruction, if a first qualified label is generated, the push plan is changed to in-depth content and discussion group invitations, and the push frequency is maintained at six times a week, and it is pushed synchronously through platform messages, emails and APP notifications.

[0185] In this embodiment, by analyzing the user's interaction time fluctuation, delay time, social influence, diversity index and time period distribution, the user's interaction behavior and influence can be deeply understood. The calculation and evaluation of the user's social activity Shy provides objective social activity evaluation results, helps the platform identify active and passive users, and adjusts the content push strategy in a targeted manner. According to the user's social activity, the push frequency and content are corrected to optimize the user experience, improve the push effect, avoid information overload, and enhance the user's acceptance of content.

[0186] Example 5, please refer to Figure 1 , further comprising a push feedback detection module, the push feedback detection module comprising a push feedback information collection unit and a push willingness calculation unit;

[0187] The push feedback information collection unit is used to collect various behavior data related to the user and the push content, establish a push feedback data set, and the push willingness calculation unit extracts the interaction characteristics between the user and the push content in the push feedback data set to calculate the user push acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i ; and push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The related calculation obtains the push content willingness Yyd;

[0188] User push acceptance rate PAP i Calculated by the following formula:

[0189]

[0190] In the formula, Js i Ts represents the number of pushed content that users accept and click in the i-th time period. i Indicates the total number of push contents received by the user in the i-th time period;

[0191] Interactive Feedback Rate (IFR) i Calculated by the following formula:

[0192]

[0193] Where Hdcs i Indicates the total number of user interactions with the pushed content in the i-th time period, including the total number of likes, comments, and shares;

[0194] User feedback delay index RFT for push content i Obtained by calculating through the following formula:

[0195]

[0196] In the formula, Fksj ij represents the feedback time after the jth push in the ith time period;

[0197] Push content relevance index PCR i Obtained by calculating through the following formula:

[0198]

[0199] In the formula, XG ij represents the relevance score between the jth push content and the user's interest in the ith time period;

[0200] And the user push acceptance rate PAP i 、User interaction feedback rate IFR i 、User feedback delay index RFT for push content i And push content relevance index PCR i , after dimensionless processing, the willingness Yyd of the push content is obtained by calculating through the following associated formula:

[0201]

[0202] In the formula, a6, a7, a8, and a9 respectively represent the user push acceptance rate PAP i 、User interaction feedback rate IFR i 、User feedback delay index RFT for push content i And push content relevance index PCR i The weight values of, which are adjusted and set by the user, 0 < a6 < 1, 0 < a7 < 1, 0 < a8 < 1, 0 < a9 < 1, and a6 + a7 + a8 + a9 = 1, b2 represents the second correction value.

[0203] The push feedback detection module includes a third evaluation unit and a second correction unit; the third evaluation unit is used to preset a fourth threshold X4, and compare and analyze the fourth threshold X4 with the push content willingness Yyd to obtain a push feedback evaluation result, including:

[0204] When the push content willingness Yyd < the fourth threshold X4, it means that the user's push feedback is not satisfied, and a first feedback anomaly label is generated;

[0205] When the push content willingness Yyd ≥ the fourth threshold X4, it means that the user's push feedback is satisfied, and a first feedback normal label is generated;

[0206] The second correction unit is used to generate a corresponding adjustment strategy according to the push feedback evaluation result, including:

[0207] If the first feedback normal tag is generated, the specific content in the second push instruction is maintained;

[0208] If the first feedback abnormal tag is generated, the second push instruction is adjusted, including:

[0209] The first low-priority content push instruction adjusts the push frequency to once a week. The push content is still short content and guiding headlines, and is pushed through a single channel of messages within the platform;

[0210] The second intermediate priority content push instruction adjusts the push frequency to 3 times a week. The push content is a combination of pictures and texts and interactive questions and answers, and is pushed simultaneously through platform messages and email push;

[0211] The third high-level priority content push instruction adjusts the push frequency to 5 times a week. The push content includes in-depth content and discussion group invitations, and is pushed simultaneously through platform messages, emails and APP notifications.

[0212] In this embodiment, the data mentioned in the specification is obtained and processed through the data record, user behavior log, social interaction monitoring and user feedback collection module of the background system;

[0213] Collect various interactive behavior data of users on pushed content to provide a detailed information basis for the evaluation of push effects. By calculating the user push acceptance rate, interactive feedback rate, feedback delay index and content relevance index, the calculation of push content willingness Yyd can comprehensively evaluate the user's acceptance and satisfaction with the pushed content. According to the evaluation results of push content willingness Yyd, adjust the push frequency and content type, optimize the user experience, and ensure the relevance and attractiveness of the pushed content. By generating feedback abnormality tags and adjusting the push strategy, it can quickly respond to user feedback on the pushed content, improve user satisfaction and the service quality of the platform.

[0214] Example 5, please refer to Figure 2 , a smart cultural science platform management method, comprising the following steps:

[0215] S1. Use web crawler tools to collect several types of cultural and popular science content from public network resources on the Internet, and establish a cultural database. After pre-processing the cultural database through a processing module, integrate and transmit it on the popular science platform;

[0216] S2, collecting the content and behavior data browsed by users on the popular science platform, and establishing a first behavior data set;

[0217] S3: Extract the personalized features, behavioral pattern features, and psychological features of the user in the first behavioral data set, and perform deep calculations using the K-means clustering algorithm and the Apriori algorithm to obtain the number of cultural classification jumps CCT of the user i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ; and classify the user culture into jump times CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i The associated calculation obtains the user's knowledge acceptance Jsd;

[0218] And evaluate the user knowledge acceptance Jsd to obtain the corresponding push level result, and generate the first push instruction according to the corresponding push level result;

[0219] S4, collecting and acquiring user social interaction data in real time and extracting social interaction features to calculate and acquire user social activity Shy, and evaluating user social activity Shy to acquire user activity evaluation results, and associating user activity evaluation results with corresponding push levels to correct the first push instruction, and acquire and execute the second push instruction;

[0220] S5. Collect various behavior data related to users and pushed content, establish a push feedback data set, and extract the interaction features between users and pushed content in the push feedback data set to calculate the user push acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i ; and push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The push content willingness degree Yyd is obtained by associated calculation; and the push content willingness degree Yyd is evaluated to generate a push feedback evaluation result, and a corresponding adjustment strategy is generated based on the push feedback evaluation result.

[0221] In this embodiment, the accuracy of content push and user experience of the platform are significantly improved through systematic data collection and analysis. First, multiple types of cultural and popular science content are collected through Web crawler tools, and integrated into the platform after preprocessing to ensure the richness and availability of content. The browsing behavior data of users on the platform are carefully analyzed to extract personalized features and behavior patterns to accurately evaluate the user's knowledge acceptance Jsd. This evaluation result is used to generate preliminary push instructions to ensure that the relevance of the pushed content is consistent with the user's interests. Secondly, the user's social interaction data is collected and analyzed in real time, the user's social activity Shy is calculated, and the push strategy is adjusted to reflect the user's interaction. This makes the pushed content not only close to the user's interests, but also adapts to the changes in their social behavior, improving the push effect and user satisfaction. By analyzing the push feedback data, the push content willingness Yyd is calculated, the acceptance of the pushed content is further evaluated, and the push strategy is adjusted according to the feedback. This ensures the high relevance and effectiveness of the pushed content, timely adjusts the push strategy to optimize the user experience, and increases the attractiveness and user stickiness of the platform content. This method realizes intelligent content management and personalized push, effectively improving the platform's operational efficiency and user satisfaction.

[0222] The threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by technicians in this field for each group of sample data; as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0223] The above formulas are obtained by collecting a large amount of data for software simulation and selecting a formula that is close to the actual value. The coefficients in the formula are set by technical personnel in this field according to actual conditions. The above is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited to this. Any technical personnel familiar with the technical field within the technical scope disclosed by the present invention, according to the technical solution and the inventive concept of the present invention, make equivalent replacement or change, which should be covered within the protection scope of the present invention.

Claims

1. An intelligent cultural science platform management system, characterized by: It includes a cultural data capture module, a processing module, a user behavior collection module, a first analysis module, a first push module and a second analysis module; The cultural data capture module is used to collect several types of cultural popular science content from public network resources on the Internet through a Web crawler tool, and to establish a cultural database. The cultural database is pre-processed by the processing module and then integrated and transmitted on the popular science platform; The user behavior collection module is used to collect the content and behavior data browsed by users on the popular science platform to establish a first behavior data set; The first analysis module is used to extract the personalized characteristics, behavioral pattern characteristics and psychological characteristics of the user in the first behavioral data set, and perform deep calculations through the K-means clustering algorithm and the Apriori algorithm to obtain the user cultural classification jump count CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i , and classify the user culture into categories and jump times CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i Associated to calculate the user knowledge acceptance Jsd; The first analysis module includes a first extraction unit and a user behavior analysis unit; The first extraction unit is used to extract the browsing behavior, clicking behavior, page jump behavior and scrolling behavior characteristics of the user on the cultural science platform from the behavior data set, and establish user personalized characteristics, behavior pattern characteristics and psychological characteristics; The user behavior analysis unit is used to perform analysis and calculation based on the user's personalized characteristics, behavior pattern characteristics and psychological characteristics to obtain: the number of user cultural classification jumps CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ; The number of user cultural category jumps CCT i Calculated by the following formula: In the formula, C j represents the number of times a user jumps from one cultural category to another in the i-th time period, and m is the number of cultural categories visited by the user in the time period; The user comprehensive scrolling index CSI i Calculated by the following formula; CSI i =S i *D i ; In the formula, S i is the page scrolling speed of the user during browsing in the i-th time period, D i is the page scrolling depth of the user's browsing process during the i-th time period; The user culture classification residence time CST i Calculated by the following formula: Where, T k represents the time the user stays on the kth cultural category page, and m is the number of cultural categories visited by the user during this time period; And the user culture classification jump number CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i The user knowledge acceptance Jsd is calculated by the following correlation formula: In the formula, n represents the number of time periods into which the user behavior data is divided, and i represents the data of the i-th time period; B i represents the frequency of the user's retrospective behavior in the i-th time period, CCT i It reflects the user's jumping behavior between different cultural categories and indicates the user's breadth of exploration in cultural categories: CSI i Reflects the user's level of interaction on the page; CST i It represents the time users spend in a certain cultural category, summarizes the degree of attention users pay to a certain cultural category, balances the time users spend in different cultural categories, and avoids excessive impact of frequent jumps or backtracking in a short period of time on the final acceptance; The first push module is used to preset a first threshold value X1 and a second threshold value X2, and compare the first threshold value X1 and the second threshold value X2 with the user knowledge acceptance degree Jsd respectively to obtain a corresponding push level result, and generate a first push instruction according to the corresponding push level result; The second analysis module is used to collect and obtain user social interaction data in real time and extract social interaction features to calculate the user's social activity Shy, and compare and analyze the social activity Shy with the third threshold value X3 to obtain the user activity evaluation result, and associate the user activity evaluation result with the corresponding push level to correct the first push instruction, obtain the second push instruction and execute it.

2. The intelligent cultural science platform management system according to claim 1, characterized in that: The cultural data capture module includes a multi-source data capture unit and a multi-type content classification unit; The multi-source data capture unit is used to collect several types of cultural popular science content from public network resources on the Internet through Web crawler tools, and establish a cultural database, and perform semantic analysis on the captured cultural data, identify the theme and keywords of the content through natural language processing (NLP) technology, and intelligently associate the relevant cultural data, and use text classification algorithms to classify the cultural database by subject, and automatically generate tags according to the content to obtain the subject classification tags. The multi-type content classification unit is used to classify the cultural database to obtain several groups of content types, and the content types include text, image, audio, video, AR / VR content, documentary materials, historical archives and oral history; The processing module is used to pre-process the cultural database, associate different types of data, including associating text content with corresponding pictures and audio, and after aggregation processing, transmit it to the science popularization platform for display.

3. The intelligent cultural science platform management system according to claim 1, characterized in that: The first pushing module includes a first evaluation unit and a first pushing unit; The first evaluation unit is used to preset a first threshold value X1 and a second threshold value X2, and compare the first threshold value X1 and the second threshold value X2 with the user knowledge acceptance Jsd respectively to obtain a corresponding push level result, including: When the user knowledge acceptance Jsd is less than the first threshold X1, the first acceptance level is obtained, indicating that the user's interest in the current content is less than 30%; When the first threshold X1≤user knowledge acceptance Jsd≤the second threshold X2, the second acceptance level is obtained, indicating that the user's interest in the current content is between 31% and 69%; When the user knowledge acceptance Jsd> the second threshold X2, the third acceptance level is obtained, indicating that the user's interest in the current content is between 70% and 100%; The first push unit is used to generate a first push instruction according to the corresponding push level result, including: Generate a first low-priority content push instruction based on the first acceptance level, including pushing a short prompt message, title guide or summary content, and set the push frequency to twice a week through a single message channel within the platform; Generate a second intermediate priority content push instruction based on the second acceptance level, including pushing short articles, picture collections or FAQs, and set the push frequency to 4 times a week, and push them simultaneously through platform messages and email push; Generate third-highest priority content push instructions based on the third acceptance level, including pushing in-depth articles, videos or expert recommendation materials, and set the push frequency to 6 times a week; and push them simultaneously through platform messages, emails and APP notifications.

4. The intelligent cultural science platform management system according to claim 1, characterized in that: The second analysis module includes a social activity data collection unit and a social interaction analysis unit; The social activity data collection unit is used to obtain user social interaction data through real-time collection; embed a tracking script in the platform page, monitor the user's click and mouse hover interaction behavior, capture real-time interaction data, and obtain user social interaction data; User social interaction data includes likes, comments, shares, discussion group participation, and friend interaction information; The social interaction analysis unit is used to extract social interaction features from user social interaction data and perform deep calculations to obtain the user interaction time fluctuation index ITVI. i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i ; User interaction time fluctuation index ITVI i Acquisition includes the following steps: Collect the user's interaction time data in the i-th time period and mark it as: i1 , L i2 ,...,L ig }where g is the number of interactive data points in the time period; Calculate the average user interaction time during this time period Where, L ij is the time of the jth interaction in the i-th time period; Calculate the variance of user interaction time during this time period In the formula, the variance It represents the average of the sum of squares of the deviations of the data points from the mean; The user interaction time fluctuation index ITVI i Calculated by the following formula: User interaction delay time index IDTI i Calculated by the following formula: In the formula, ΔYc ij is the latency of the jth interaction in the i-th time period, the time from when the user initiates the interaction to when the response is received; g is the number of interaction data points in the time period; The user social influence index S11 i Calculated by the following formula: In the formula, E i is the number of times the user interaction content is cited or forwarded by other users in the i-th time period; R i is the total number of user interactions in the i-th time period; User Interaction Diversity Index IDI i Calculated by the following formula: In the formula, F ij is the frequency of the jth social interaction type in the i-th time period, and the social interaction types include comments, likes, and shares; T ij is the total number of interactions of the jth type in the ith time period; e is the total number of interaction types; The user interaction period distribution ITDI i Calculated by the following formula: In the formula, Entropy(H i ) is the entropy value of the user interaction time distribution in the i-th time period, H max is the maximum entropy value.

5. The intelligent cultural science platform management system according to claim 4 is characterized by: The second analysis module also includes a user social activity calculation unit, a second evaluation unit and a first correction unit; The user social activity calculation unit is used to calculate the user interaction time fluctuation index ITVI i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i After dimensionless processing, the user's social activity Shy is calculated using the following associated formula: Where a1, a2, a3, a4 and a5 represent the user interaction time fluctuation index ITVI i , User Interaction Delay Time Index IDTI i , User social influence index Sll i , User Interaction Diversity Index IDI i ITDI distribution during user interaction period i The weight value of b1 represents the first correction constant value; The second evaluation unit is used to preset a third threshold value X3, and compare the third threshold value X3 with the user's social activity Shy to obtain a user activity evaluation result, including: When the user's social activity Shy is less than the third threshold value X3, it means that the user's social activity is unqualified, and a first unqualified label is generated, indicating that the user is in a passive state in social interaction; When the user's social activity Shy ≥ the third threshold value X3, it means that the user's social activity is qualified, and a first qualified label is generated, indicating that the user is in an active state in social interaction; The first correction unit is used to associate the user activity evaluation result with the corresponding push level to correct the first push instruction to obtain the second push instruction, including: If a user who receives the first low-priority content push instruction generates the first unqualified label, the push frequency setting is changed to once a week; if a user who receives the second medium-priority content push instruction generates the first unqualified label, the push frequency setting is changed to three times a week; if a user who receives the third high-priority content push instruction generates the first unqualified label, the push frequency setting is changed to five times a week; For users who receive the first low-priority content push instruction and generate the first qualified tag, the push plan will be short content and guiding titles, and the push frequency will be maintained at 2 times a week, which will be pushed through a single channel of platform messages; for users who receive the second intermediate-priority content push instruction and generate the first qualified tag, the push plan will be a combination of pictures and texts and interactive questions and answers, and the push frequency will be maintained at 4 times a week, which will be pushed simultaneously through platform messages and emails; for users who receive the third-highest-priority content push instruction and generate the first qualified tag, the settings will be changed to in-depth content and discussion group invitations, and the push frequency will be maintained at 6 times a week, which will be pushed simultaneously through platform messages, emails and APP notifications.

6. The intelligent cultural science platform management system according to claim 1, characterized in that: It also includes a push feedback detection module, which includes a push feedback information collection unit and a push willingness calculation unit; The push feedback information collection unit is used to collect various behavior data related to the user and the push content, establish a push feedback data set, and the push willingness calculation unit extracts the interaction characteristics between the user and the push content in the push feedback data set to calculate the user push acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i ; and push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The related calculation obtains the push content willingness Yyd; The user push acceptance rate PAP i Calculated by the following formula: In the formula, Js i Ts represents the number of pushed content that users accept and click in the i-th time period. i Indicates the total number of push contents received by the user in the i-th time period; Interactive Feedback Rate (IFR) i Calculated by the following formula: Where Hdcs i Indicates the total number of user interactions with the pushed content in the i-th time period, including the total number of likes, comments, and shares; The user's feedback delay index RFT for the pushed content i Calculated by the following formula: In the formula, Fksj ij Indicates the feedback time after receiving the jth push in the i-th time period; The push content relevance index PCR i Calculated by the following formula: In the formula, XG ij Indicates the relevance score between the jth pushed content and the user's interests in the i-th time period; And push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i After dimensionless processing, the push content willingness Yyd is calculated through the following related formula: Where a6, a7, a8 and a9 represent the user push acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The weight value of b2 represents the second correction constant value.

7. The intelligent cultural science platform management system according to claim 6 is characterized by: The push feedback detection module includes a third evaluation unit and a second correction unit; the third evaluation unit is used to preset a fourth threshold value X4, and compare and analyze the fourth threshold value X4 with the push content willingness Yyd to obtain a push feedback evaluation result, including: When the push content willingness Yyd is less than the fourth threshold value X4, it indicates that the user is dissatisfied with the push feedback, and a first feedback abnormality label is generated; When the push content willingness Yyd ≥ the fourth threshold X4, it means that the user is satisfied with the push feedback, and a first feedback normal label is generated; The second correction unit is used to generate a corresponding adjustment strategy according to the push feedback evaluation result, including: If the first feedback normal tag is generated, the specific content in the second push instruction is maintained; If the first feedback abnormal tag is generated, the second push instruction is adjusted, including: The first low-priority content push instruction adjusts the push frequency to once a week. The push content is still short content and guiding headlines, and is pushed through a single channel of messages within the platform; The second intermediate priority content push instruction adjusts the push frequency to 3 times a week. The push content is a combination of pictures and texts and interactive questions and answers, and is pushed simultaneously through platform messages and email push; The third high-level priority content push instruction adjusts the push frequency to 5 times a week. The push content includes in-depth content and discussion group invitations, and is pushed simultaneously through platform messages, emails and APP notifications.

8. An intelligent culture and science popularization platform management method, applied to an intelligent culture and science popularization platform management system according to any one of claims 1 to 7, characterized in that: The following steps are involved: S1. Using a Web crawler tool to publicly disclose network resources from the Internet, collect several types of cultural and popular science content, and establish a cultural database. After pre-processing the cultural database through the processing module, it is integrated and transmitted on the popular science platform; S2, collecting the content and behavior data browsed by users on the popular science platform, and establishing a first behavior data set; S3: Extract the personalized features, behavioral pattern features, and psychological features of the user in the first behavioral data set, and perform deep calculations using the K-means clustering algorithm and the Apriori algorithm to obtain the number of cultural classification jumps CCT of the user i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i ; and classify the user culture into jump times CCT i , User Comprehensive Rolling Index CSI i and user culture classification residence time CST i The associated calculation obtains the user's knowledge acceptance Jsd; And evaluate the user knowledge acceptance Jsd to obtain the corresponding push level result, and generate the first push instruction according to the corresponding push level result; S4, collecting and acquiring user social interaction data in real time and extracting social interaction features to calculate and acquire user social activity Shy, and evaluating user social activity Shy to acquire user activity evaluation results, and associating user activity evaluation results with corresponding push levels to correct the first push instruction, and acquire and execute the second push instruction; S5. Collect various behavior data related to users and pushed content, establish a push feedback data set, and extract the interaction features between users and pushed content in the push feedback data set to calculate the user push acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i ; and push the user acceptance rate PAP i , user interaction feedback rate IFR i , User feedback delay index RFT for pushed content i PCR index of push content relevance i The push content willingness degree Yyd is obtained by associated calculation; and the push content willingness degree Yyd is evaluated to generate a push feedback evaluation result, and a corresponding adjustment strategy is generated based on the push feedback evaluation result.

Images