Regional culture transmission-based feasibility evaluation system

Abstract

The invention relates to the technical field of cultural transmission, in particular to a regional cultural transmission-based feasibility evaluation system, which comprises a data integration and formatting module, a similarity matrix generation module, a prediction analysis module, a behavior data analysis module and a transmission influence evaluation module. According to the method, the user similarity matrix is accurately constructed and the content is effectively recommended by carefully integrating and formatting the user activity data, utilizing normalization processing and Pearson's correlation coefficient calculation, and the content is effectively recommended by analyzing the popularity degree of the unscored topics and the user behavior trend. According to the method, the popularization of scientific culture activities is enabled to more accurately adapt to specific requirements of different regions and groups, the pertinence and the effect prediction capability of culture transmission are enhanced, resource allocation and strategy formulation are optimized, and the participation degree and culture influence of the activities are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cultural dissemination, and particularly to a feasibility evaluation system for regional cultural dissemination. Background Art

[0002] The technical field of cultural dissemination mainly involves using various media and platforms to promote and spread cultural knowledge and information, integrating traditional cultural research with modern information technology. The focus is on how to effectively convey cultural values and educational content to the public or specific groups through digital means, including aspects such as content creation, editing, distribution, and feedback collection. With the progress of technology, the cultural dissemination field increasingly relies on the Internet, social media, and other digital tools to improve its efficiency and breadth, while also exploring how to more precisely target specific cultural needs and responses through data analysis and regional research.

[0003] Among them, the feasibility evaluation system for regional cultural dissemination refers to a system that uses data analysis technology to evaluate the feasibility of scientific and cultural dissemination activities in a specific region. By analyzing factors such as the cultural background, educational level, and public interest of the region, it evaluates the promotion possibility and expected effect of scientific-related activities or content. Its uses include helping cultural and educational institutions and related organizations formulate science popularization plans that better conform to regional characteristics and needs, optimizing resource allocation, and enhancing the influence and participation of scientific and cultural activities.

[0004] In the prior art, when dealing with user data and feedback, in-depth behavior analysis is often not carried out, resulting in inaccurate docking between cultural dissemination activities and the needs of the target audience. The lack of precise user similarity analysis and personalized content recommendation systems often leads to generally insufficient promotion effects of cultural content and fails to fully utilize the potential of digital tools. In addition, the prior art often ignores the in-depth study of regional cultural characteristics and fails to optimize communication strategies according to regional characteristics, restricting the wide acceptance and participation of scientific and cultural activities. For example, the lack of analysis of the behavior trends of specific groups may lead to the promotion plan being unable to effectively stimulate public interest, affecting the far-reaching effect of cultural dissemination, making the traditional methods show obvious inefficiency in resource allocation and effect evaluation, and reducing the popularization and educational effect of science and culture. Summary of the Invention

[0005] The purpose of the present invention is to solve the deficiencies existing in the prior art, and to propose a feasibility evaluation system for regional cultural dissemination.

[0006] To achieve the above purpose, the present invention adopts the following technical solutions: The feasibility evaluation system for regional cultural dissemination includes: The data integration and formatting module extracts user video click records, viewing durations, and rating data, organizes and enters the video click records, viewing durations, and rating data into the database, creates data rows for each user and topic, checks the integrity and accuracy of the data, and generates user activity characteristic data; The similarity matrix generation module statistically analyzes the user rating patterns based on the user activity characteristic data, performs normalization processing on the rating data of each user respectively, calculates the correlation coefficient, compares the rating vectors between different users one by one according to the calculation results, analyzes the rating similarity of all users for the corresponding topics, and constructs and generates a similar user analysis matrix; The prediction analysis module analyzes the potential popularity of each unrated topic based on the similar user analysis matrix, makes rating predictions for each topic, integrates the user preference patterns, and outputs the predicted topic acceptance indicators; The behavior data analysis module groups the user behavior data according to the predicted topic acceptance indicators, conducts trend analysis on each group of data, identifies common and abnormal behavior patterns, reveals the behavior characteristics of each group according to the identification results, and corresponds them to the regions to obtain the regional behavior trend analysis results; The dissemination impact assessment module measures the acceptance of the topic in the target region based on the regional behavior trend analysis results, evaluates the actual effect of content dissemination, determines the regional cultural adaptation degree, and generates the regional cultural dissemination effect assessment results.

[0007] As a further solution of the present invention, the steps for obtaining the user activity characteristic data are as follows: Collect user click records, viewing durations, and rating data, filter all viewing behavior data of the same user within the same time window, sort them according to the user identifier, topic identifier, and timestamp, and use the formula: ; Calculate the user's effective viewing duration , and generate user effective viewing duration data, where represents the residence duration of the user's th viewing behavior, represents the rating weight of the th viewing behavior, represents the total number of viewing behaviors; Based on the user effective viewing duration data, match the user's click behavior and rating information, establish the viewing behavior records of the user within the target time window, classify and organize the user's viewing duration, click times, and rating results, and construct user viewing behavior data rows according to the user identifier and topic identifier; Screen abnormal data and error data according to the user viewing behavior data line, match the existing user behavior patterns, verify the data integrity, create corresponding user activity feature data for each user and topic, and enter it into the database.

[0008] As a further solution of the present invention, the calculation steps of the correlation coefficient are as follows: Based on the user activity feature data, extract the rating records of each user, collect the users with complete rating data, calculate the average rating and the standard deviation of the ratings of each user, and generate user rating pattern data; According to the user rating pattern data, collect the minimum and maximum values of the rating data of each user, map the user rating data to the standard range, calculate the normalized rating value, and organize and output the user normalized rating data; Combined with the user normalized rating data, based on the covariance and standard deviation of the rating vectors, use the formula: ; Calculate the correlation coefficient between user rating vectors , and obtain the user rating correlation coefficient data, where represents user and user 's rating correlation coefficient, represents the normalized rating value of user on the th rating item, represents the normalized rating value of user on the th rating item, represents the average rating of user , represents the average rating of user , represents the total number of rating items.

[0009] As a further solution of the present invention, the steps for obtaining the similar user analysis matrix are as follows: According to the user rating correlation coefficient data, combine the preset threshold to screen the user pairs with the correlation coefficient greater than the preset threshold, and use the formula: ; Calculate the rating vector distance between user and user , and obtain the user rating vector similarity data, where represents the normalized rating value of user on the th rating item, represents the normalized rating value of user on the The normalized scoring value on each scoring item represents the total number of scoring items; According to the user scoring vector similarity data, analyze the scoring consistency of all users for each topic, compare the scoring vectors of all users one by one, analyze the overlap degree of the scoring vectors on a specific topic, and obtain the scoring similarity data of users for the topic; Based on the scoring similarity data of users for the topic, with the user identifier as the row index, the topic identifier as the column index, and the scoring similarity of the user for the topic as the filling matrix element, construct and output the similar user analysis matrix.

[0010] As a further solution of the present invention, the steps for obtaining the predicted topic acceptance index are as follows: Based on the similar user analysis matrix, screen the topics that each user has not scored, extract the scoring data of similar users for the corresponding topics, combine the scoring behaviors of similar users to identify the scoring bias, and establish the data of the potential popularity of the topics; According to the data of the potential popularity of the topics, perform scoring prediction on each topic, analyze the interest changes of the target user on different topics, compare the historical scoring records and topic preferences of the target user, and combine the scoring fluctuation analysis of the user for the topic to generate the topic scoring prediction data; Based on the topic scoring prediction data, analyze the scoring tendencies of different users on each topic, classify the scoring characteristics of different user groups, determine the user interest distribution, summarize the overall scoring trend, and output the predicted topic acceptance index.

[0011] As a further solution of the present invention, the steps for identifying the behavior pattern are as follows: Based on the predicted topic acceptance index, collect the scoring data of users on different topics, classify them according to the scoring characteristics, scoring stability and scoring range of the users, calculate the average value and the range of scoring changes of the scores of different users, and group the users according to the similarity of the scoring patterns, and output the user behavior grouping data; Based on the user behavior grouping data, analyze the scoring trend of each group, extract the average scoring value of the group at consecutive time points, and use the formula: ; Calculate the scoring change rate from to within the time period of the group, and obtain the group behavior trend data, where represents the average score of the group at time , represents the average score of the group at time The average score, represents a time interval; Based on the group behavior trend data, compare the scoring patterns within the group, screen out the common behavior patterns with stable scoring changes, extract the individuals with abnormal scoring changes, analyze the scoring behaviors of abnormal users, summarize the scoring deviation characteristics, and organize and output the data of common and abnormal behavior patterns.

[0012] As a further solution of the present invention, the steps for obtaining the regional behavior trend analysis result are as follows: According to the data of the common and abnormal behavior patterns, analyze the scoring trend characteristics of each group, extract the stability of the scoring changes within the group, the scoring deviation situation and the scoring fluctuation range, analyze the degree of consistency of the user behaviors within the group, compare the distribution of the scoring preferences of the users within the group, and summarize and generate the group behavior characteristic data; Based on the group behavior characteristic data, match the geographical region information of the group users, extract the behavior patterns of the user groups in different regions, compare the scoring trends of the users within the region, count the change rate of the region users' scores, analyze the stability of the user behavior characteristics within the region, evaluate the scoring differences between different regions, and obtain the region user behavior data; Combined with the region user behavior data, analyze the behavior characteristics of different regions, summarize the scoring preferences of the region user groups, match the geographical characteristics and user activity levels of the regions, summarize the development trends of the user behaviors in each region, and output the regional behavior trend analysis result.

[0013] As a further solution of the present invention, the steps for obtaining the evaluation result of the regional cultural dissemination effect are as follows: Based on the regional behavior trend analysis result, extract the scoring data of all users in the target region, screen out the scoring records of the target topic, and arrange them in chronological order, calculate the regional average score, score variance, skewness and kurtosis of the scoring distribution of the topic, and summarize and obtain the acceptance data of the target region users for the topic; Based on the acceptance data of the target region users for the topic, count the scoring change situations in different time periods of the target region, calculate the growth rate and decay rate of the topic score, and analyze the change range of the score over time, evaluate the user penetration of the content dissemination, summarize the dissemination effect of the topic in different time ranges, and obtain the content dissemination effect evaluation result; Based on the content dissemination effect evaluation result, compare the dissemination fitness of different topics in the target region, combine the cultural characteristics of the target region, analyze the preferences of the region users for different topics, determine the cultural adaptability and summarize the cultural adaptation characteristics of the target region, and organize and output the evaluation result of the regional cultural dissemination effect.

[0014] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In the present invention, through the meticulous integration and formatting of user activity data, using normalization processing and Pearson correlation coefficient calculation, a user similarity matrix is precisely constructed to effectively recommend content. By analyzing the popularity of unrated topics and user behavior trends, the promotion of scientific and cultural activities more accurately adapts to the specific needs of different regions and groups, not only enhancing the pertinence and effect prediction ability of cultural dissemination, but also optimizing resource allocation and strategy formulation, greatly improving the participation rate and cultural influence of activities. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 is the system flow chart of the present invention; Figure 2 is the acquisition flow chart of user activity feature data of the present invention; Figure 3 is the calculation flow chart of the correlation coefficient of the present invention; Figure 4 is the acquisition flow chart of the similar user analysis matrix of the present invention; Figure 5 is the acquisition flow chart of the predicted topic acceptance index of the present invention; Figure 6 is the recognition flow chart of the behavior pattern of the present invention; Figure 7 is the acquisition flow chart of the regional behavior trend analysis result of the present invention; Figure 8 is the acquisition flow chart of the regional cultural dissemination effect evaluation result of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0017] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc. is based on the orientation or positional relationship shown in the accompanying drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as limiting the present invention. In addition, in the description of the present invention, "a plurality" means two or more unless otherwise specifically defined.

[0018] Please refer to Figure 1 , the regional cultural dissemination feasibility evaluation system based on includes: The data integration and formatting module extracts the user video click records, viewing durations, and rating data, organizes and enters the video click records, viewing durations, and rating data into the database, creates data rows for each user and topic, verifies the integrity and accuracy of the data, and generates user activity characteristic data; The similarity matrix generation module, based on the user activity characteristic data, statistically analyzes the user rating patterns, performs normalization processing on the rating data of each user respectively, calculates the correlation coefficient, compares the rating vectors between different users one by one according to the calculation results, analyzes the rating similarity of all users for the corresponding topics, and constructs and generates a similar user analysis matrix; The predictive analysis module, based on the similar user analysis matrix, analyzes the potential popularity of each unrated topic, predicts the ratings of each topic, integrates the user preference patterns, and outputs the predicted topic acceptance indicators; The behavior data analysis module groups the user behavior data according to the predicted topic acceptance indicators, conducts trend analysis on the data of each group, identifies common and abnormal behavior patterns, reveals the behavior characteristics of each group according to the identification results, and corresponds them to the regions to obtain the regional behavior trend analysis results; The communication impact assessment module, based on the regional behavior trend analysis results, measures the acceptance of the topic in the target region, evaluates the actual effect of the content communication, determines the regional cultural adaptation degree, and generates the regional cultural communication effect assessment results.

[0019] The user activity characteristic data includes user identity identification, activity timestamps, and topic interaction details. The similar user analysis matrix includes user association degree, topic similarity index, and similarity rating. The predicted topic acceptance indicators specifically refer to topic prediction scores, popularity assessment results, and user group preference analysis records. The regional behavior trend analysis results include behavior frequency analysis results, abnormal behavior indicators, and group preference trends. The regional cultural communication effect assessment results are specifically regional acceptance assessment, cultural adaptation degree analysis results, and communication effect levels.

[0020] Please refer to Figure 2 , the steps for obtaining the user activity characteristic data are as follows: Collect the user click records, viewing durations, and rating data, filter all the viewing behavior data of the same user within the same time window, sort them according to the user identification, topic identification, and timestamp, and use the formula: ; Calculate the user's effective viewing duration , and generate the user's effective viewing duration data, where represents the residence duration of the user on the th viewing behavior, and represents the The scoring weight of viewing behavior represents the total number of viewing behaviors; Viewing duration : By monitoring the start and end times of each video viewing by the user through video playback, calculate the duration of each viewing. For example, if the start time of the user's viewing of the first video is 10:00 and the end time is 10:05, then minutes.

[0021] Scoring weight : Based on the user's rating of each video, usually between 1 and 5 points. Assume the user's rating of the first video is 4 points, then .

[0022] Total number of viewing behaviors : Count the number of times the user views videos within a specific time window. For example, if the user views 10 videos within a week, then .

[0023] Suppose a user views the following videos within a week: ; Calculation process: Calculate the numerator part: ; Calculate the denominator part: ; Calculate the user's effective viewing duration : ; The result shows that the average effective viewing duration of the user within the selected time window is about 7.17 minutes, reflecting the overall interest level of the user in the viewed videos and helping to further analyze the user's viewing preferences.

[0024] Based on the user's effective viewing duration data, match the user's click behavior and rating information, establish the viewing behavior record of the user within the target time window, classify and organize the user's viewing duration, click times and rating results, and construct the user viewing behavior data row according to the user identification and topic identification; Based on the user's effective viewing duration data, match the user's click behavior and rating information. First, collect all the viewing records of the user within a specific time window, including the start time, end time, identifier of the video watched each time, and the user's rating of the video. Store this data in a structured data table. Then, calculate the duration of each viewing, which is the end time minus the start time, to obtain the viewing duration. Combine the viewing duration with the user rating to generate a dataset containing user identifiers, video identifiers, viewing duration, and ratings. Next, group the data by user identifier and video identifier, and summarize the total viewing duration and average rating of each user for each video. Finally, store the organized data in the user viewing behavior table in the database for subsequent analysis and processing.

[0025] Screen abnormal data and error data according to the user viewing behavior data rows, match the existing user behavior patterns, check the data integrity, create corresponding user activity feature data for each user and topic, and enter it into the database; Call the user viewing behavior data rows, check the data integrity, screen abnormal data and error data. First, extract the data from the user viewing behavior table in the database, and check whether each record contains complete user identifier, video identifier, viewing duration, and rating information. For records missing any key fields, mark them as incomplete data. Then, analyze the value ranges of the viewing duration and rating, set reasonable upper and lower threshold values, and mark the records outside the threshold range as abnormal data. Next, check the data consistency. For example, ensure that the sum of multiple viewing records of the same user for the same video is consistent with the summarized data. When inconsistencies are found, mark them as error data. Finally, summarize the marked incomplete, abnormal, and error data to generate a data quality report for data maintenance personnel to further process and correct.

[0026] Please refer to Figure 3 , and the calculation steps of the correlation coefficient are as follows: Based on the user activity feature data, extract the rating records of each user, collect the users with complete rating data, and calculate the average rating and rating standard deviation of each user to generate user rating pattern data; According to the user activity feature data, extract the rating records of each user. First, read the user rating data from the database, filter out the users with complete rating records, and remove the missing or abnormal rating data. Then, calculate the average rating and rating standard deviation of each user, respectively count the rating distribution of the users, check the concentration and dispersion of the ratings, compare the average rating and standard deviation with the historical rating patterns, identify the abnormal situations of the rating patterns, and judge the stability of the rating trends. Finally, store the user average rating, rating standard deviation, and rating trend data to generate user rating pattern data.

[0027] According to the user rating pattern data, collect the minimum and maximum values of the rating data for each user, map the user rating data to the standard range, calculate the normalized rating value, and organize and output the user normalized rating data; During the calculation of the normalized rating value, use the formula: ; Perform the calculation, where represents the normalized rating value of user , represents the original rating value of user , represents the minimum value among all ratings, represents the maximum value among all ratings; Suppose the rating data of a certain user is . First, calculate the minimum and maximum values of all rating data: , . Then, perform the normalization calculation for each rating value.

[0028] For the rating : ; For the rating : ; For the rating : ; For the rating : ; For the rating : ; The results show that the user rating data has been normalized to the range of 0 to 1, enabling the comparison of user data in different rating ranges on the same scale, thus ensuring the balance and consistency of the rating data.

[0029] Combined with the user normalized rating data, based on the covariance and standard deviation of the rating vectors, use the formula: ; Calculate the correlation coefficient between user rating vectors to obtain the user rating correlation coefficient data, where represents the rating correlation coefficient between user and user , represents user at the The normalized scoring value on the -th scoring item, which represents the user 's normalized scoring value on the -th scoring item, represents the average score of the user ; represents the average score of the user ; represents the total number of scoring items; Suppose the normalized score of user a is , and the normalized score of user b is ; ; Calculate the numerator part of the covariance: ; ; ; Calculate the denominator part of the standard deviation: ; ; ; ; ; Finally, calculate the Pearson correlation coefficient: ; The result shows that the scoring patterns of user a and user b are highly similar, and the correlation coefficient between their scoring behaviors is 0.943, indicating that the scoring trends of the two users for the same items are relatively consistent, which can be used for user similarity analysis and recommendation system optimization.

[0030] Please refer to Figure 4 for the steps to obtain the similar user analysis matrix: According to the user scoring correlation coefficient data, filter the user pairs with correlation coefficients greater than the preset threshold in combination with the preset threshold, using the formula: ; Calculate the distance between the scoring vectors of user and user to obtain the user scoring vector similarity data, where represents the normalized scoring value of user on the -th scoring item, represents the user The normalized rating value on the th rating item, represents the total number of rating items; Assume that the normalized ratings of user c and user d on 5 topics are as follows: ; Calculate the Euclidean distance First, calculate the square of the rating difference for each topic: ; ; ; ; ; Then, sum up the squared values: ; Take the square root of the result to obtain the Euclidean distance: ; The result shows that the Euclidean distance between the rating vectors of user c and user d is 0.332, which measures the rating similarity between the two users. The smaller the value, the more similar the ratings of the two users. Conversely, a larger value indicates a greater difference in the rating patterns. In practical applications, if the rating distance threshold for similar users is set to 0.5, then the rating distance of 0.332 between user c and user d is less than 0.5. Therefore, it can be determined that the rating patterns of these two users are relatively similar and they can be classified as potential similar users.

[0031] According to the user rating vector similarity data, analyze the rating consistency of all users for each topic. Compare the rating vectors of all users one by one, analyze the overlap degree of the rating vectors on a specific topic, and obtain the user rating similarity data for the topic; Call the user rating vector similarity data. First, traverse the rating vectors of all users to ensure that the dimensions of the rating vectors of each user are consistent and sort them according to the same rating items, so that the data is comparable between different users. Then, for each pair of user rating vectors, calculate the rating difference for each topic and compare it according to the set similarity threshold. If the difference is within the threshold range, mark it as rating similar; otherwise, mark it as rating different. Next, calculate the rating consistency score of all users on each topic, using the weighted summation method, and calculate the overall rating consistency of each topic according to the contribution ratio of the rating vector similarity. Archive and store the calculation results in the rating consistency table to ensure that the data is available for subsequent analysis and call. Finally, aggregate the rating similarity data of all users on different topics to lay a foundation for further analysis of the rating relationship between users.

[0032] Based on the similarity data of users' ratings on topics, with user identifiers as row indices, topic identifiers as column indices, and the similarity of users' ratings on the topic as the elements filling the matrix, construct and output a similar user analysis matrix; Call the similarity data of users' ratings on topics. First, create a matrix structure with user identifiers as row indices and topic identifiers as column indices, and initialize each element in the matrix to zero. Then, traverse the similarity data of users' ratings on topics. For each pair of users, check their rating similarities on each topic. If the rating similarity between two users on a certain topic exceeds the set matching threshold, update the corresponding element value in the matrix and assign the corresponding rating similarity weight. Next, perform normalization processing on all elements in the matrix to ensure that all similarity values are within a unified numerical range for subsequent calculation and analysis. After normalization, format the matrix so that it can be associated with user feature data. Finally, output the processed similar user analysis matrix, which can be used in subsequent application scenarios such as user relationship mining, interest matching, and recommendation strategy optimization.

[0033] Please refer to Figure 5 , the steps to obtain the predicted topic acceptance index are as follows: Based on the similar user analysis matrix, filter out the topics that each user has not rated, extract the rating data of similar users on the corresponding topics, and combine the rating behaviors of similar users to identify rating biases and establish topic potential popularity data; Call the similar user analysis matrix to analyze the potential popularity of each unrated topic. First, filter out the topics that each user has not rated to ensure that all unrated topics can match the rating data of at least one similar user, and exclude those topics that cannot obtain rating references from similar users. Then, for each unrated topic, find the group of similar users of the user, determine the selection criteria for similar users to ensure that there is a certain correlation in the rating habits among similar users. Next, collect the rating records of similar users on the topic, calculate the statistical characteristics of the rating data, including the mean, median, rating variance, and rating distribution range of the ratings, identify possible outliers in the rating data, and remove the data points with ratings significantly deviating from other ratings to prevent extreme ratings from affecting the calculation results. Subsequently, calculate the weighted average of the rating data, combine the weights of similar users' ratings with the user similarity so that the rating contribution degree is proportional to the similarity between users, and ensure that the rating popularity can fully reflect the possible rating trend of the target user. Finally, organize the potential popularity data of all topics, store it in a structured data table, and perform necessary data annotation for subsequent analysis and invocation.

[0034] Based on the data of the potential popularity of topics, a score prediction is made for each topic, the interest changes of target users on different topics are analyzed, the historical score records and topic preferences of target users are compared, and combined with the analysis of the score fluctuations of users on topics, topic score prediction data is generated; Call the data of the potential popularity of topics and make a score prediction for each topic. First, extract the historical score records of target users, calculate the average value, score range and score fluctuation of their scores, and judge whether the users have score preferences, such as whether they often give high scores or low scores, and whether they tend to give higher scores to specific types of topics. Then, analyze the score distribution of target users, calculate the score variance of users on different types of topics, determine the stability of users' scores, identify the degree of bias of users' scores between different topics. Next, compare the scoring habits of target users with the scoring behaviors of similar users, calculate the potential score differences of target users on unrated topics, and correct the data of the potential popularity of topics to make it more compatible with the scoring style of target users, avoiding the deviation of score prediction values from the scoring habits of target users due to directly citing the scores of similar users. Subsequently, adjust the score prediction values to ensure that the predicted scores are reasonable and conform to the scoring patterns of users. Finally, organize all the predicted score data, store it classified by users and topics, and provide data support for subsequent analysis and recommendation.

[0035] Based on the topic score prediction data, analyze the scoring tendencies of different users on each topic, classify the scoring characteristics of different user groups, determine the user interest distribution, summarize the overall scoring trend, and output the predicted topic acceptance index; Call the topic score prediction data and integrate the user preference patterns. First, classify the score data of users according to the score range and scoring trend, extract the scoring tendencies of users on different types of topics, and calculate the key indicators of users' scoring behaviors, including the stability, score difference, scoring change frequency and scoring consistency of users' scores, and identify the scoring characteristics of users on different types of topics. Then, match the predicted score data with the historical score distribution of users, analyze the degree of interest of users on each topic, identify the types of topics that users may be interested in, and analyze the acceptance of users for specific topics based on historical score data. Next, calculate the scoring trends of all users on all topics, analyze the score differences of different user groups on each topic, and identify which topics are more popular among the overall user group and which topics have lower scores among specific user groups. Subsequently, normalize the score prediction data to ensure that the score data of different users are compared on the same scale, and adjust the score acceptance data to make it more conform to the user scoring trend. Finally, organize the acceptance indicators of all users, file them according to user interest categories, and output the predicted topic acceptance index for subsequent analysis and recommendation.

[0036] Please refer toFigure 6 , the steps for identifying the behavior pattern are as follows: Based on the predicted topic acceptance index, collect the rating data of users on different topics, classify them according to the rating characteristics, rating stability, and rating range of users, calculate the average value and the range of rating changes of different users' ratings, group users according to the similarity of rating patterns, and output the user behavior grouping data; According to the predicted topic acceptance index, extract the rating data of users on different topics. First, screen out the historical rating records of users to ensure that the rating data of all users contain complete time series information and can match the corresponding topic classifications. Then, classify the user data according to the changing trend of ratings, calculate the rating change rate of each user, and smooth the changing trend of ratings to identify potential rating patterns. Next, analyze the amplitude of rating fluctuations, calculate the rating stability of different users in each time period, and based on the standard deviation and distribution of ratings, determine the rating stability category of users. Subsequently, according to the similarity of rating patterns, group users according to the rating trend, rating stability, and rating fluctuation range to ensure that users in each group have similar rating habits. Finally, organize the user behavior grouping data and label the characteristics of each group for subsequent analysis.

[0037] Based on the user behavior grouping data, analyze the rating trend of each group, extract the average rating value of the group at consecutive time points, and use the formula: ; Calculate the rating change rate of the group within the time period to to obtain the group behavior trend data, where represents the average rating of group at time , represents the average rating of group at time , represents the time interval; The average ratings of a certain group at 4 time points are collected as follows: The average ratings of a certain group at 4 time points are collected as follows: ; To calculate the rating change rate, first, calculate the rating change rate of time compared to : ; Then, calculate the rating change rate of time compared to : ; Finally, calculate the time Compared to Rating change rate: ; The results show that the score trend of this group in the first two time points was on the rise, with the score increasing by 0.3 points in each time period, while at the last time point, the score showed a downward trend, decreasing by 0.2 points. This result reflects the dynamic changes in the group score and helps to further analyze the behavioral trends.

[0038] Based on group behavior trend data, compare the rating patterns within the group, screen common behavior patterns with stable rating changes, extract individuals with abnormal rating changes, analyze the rating behavior of abnormal users, summarize the rating deviation characteristics, and organize and output common and abnormal behavior pattern data; Combined with the group behavior trend data, the scoring patterns within the group are compared. First, the scoring sequences of users in each group are extracted, and the change range and stability of the scores are analyzed to ensure that the data is complete and can match the scoring trends of each time period. Then, the similarities between the scoring patterns are calculated to identify the patterns of relatively stable score changes within the group. Groups with small score changes and stable score trends are screened out as representatives of common behavior patterns. Next, groups with drastic score fluctuations are analyzed to identify users with high frequency of score changes or sudden score changes. Individuals with scores that deviate greatly from the average are screened, and the degree of deviation of their scores is calculated to determine whether they are abnormal behaviors. Subsequently, common and abnormal scoring patterns are classified and sorted according to the characteristics of score changes. Finally, the scoring pattern characteristics within the group are summarized, and common behavior pattern and abnormal behavior pattern data are output for subsequent analysis.

[0039] See also Figure 7 , the steps to obtain the regional behavior trend analysis results are: Based on the common and abnormal behavior pattern data, analyze the rating trend characteristics of each group, extract the stability of the rating changes within the group, the rating deviation and the rating fluctuation range, analyze the consistency of user behavior within the group, compare the rating preference distribution of users within the group, and summarize and generate group behavior characteristic data; According to the data of common and abnormal behavior patterns, analyze the scoring trend characteristics of each group. First, extract the scoring records of all users within the group, arrange them in chronological order to ensure that the scoring data covers the complete time period, and remove duplicate or invalid scoring data to ensure the continuity and analyzability of the data. Then, calculate the stability of the scores within the group, including the variance of user scores, the range of score distribution, and the rate of score change, identify the user groups with large score fluctuations, and calculate the deviation degree of the average score of each user from the group average. Next, compare the scoring preferences of users within the group, analyze the scoring similarity of different users on specific topics, and identify the user groups with convergent or significantly different scoring patterns. Subsequently, summarize the group behavior characteristics based on the score stability, score deviation, and score change characteristics. Finally, organize and output the group behavior characteristic data to provide support for subsequent regional matching analysis.

[0040] Based on the group behavior characteristic data, match the geographical region information of the group users, extract the behavior patterns of the user groups in different regions, compare the scoring trends of the users within the regions, count the rate of change of the scores of the regional users, analyze the stability of the behavior characteristics of the users within the regions, evaluate the scoring differences between different regions, and obtain the regional user behavior data. According to the group behavior characteristic data, match the geographical region information of each group of users. First, extract the location information of all users within the group to ensure that all users can be matched to their respective regions, classify the user location information, and establish the corresponding relationship between users and regions. Then, count the scoring characteristics of different groups within each region, calculate the average scoring level, the rate of change of scores, and the score deviation of the users within each region to ensure that the data of different regions can be compared on the same scale. Next, analyze the behavior patterns of the regional user groups, identify the scoring trends of the users in each region, calculate the range of score distribution and the variance of scores within different regions, and evaluate the stability and consistency of the scoring patterns of the users in each region. Subsequently, compare the regional user scoring patterns with the scoring patterns of each group to identify the groups with a high degree of scoring pattern matching, and summarize the scoring characteristics of different regions. Finally, organize and output the regional user behavior data for subsequent regional trend analysis.

[0041] Combined with the regional user behavior data, analyze the behavior characteristics of different regions, summarize the scoring preferences of the regional user groups, match the geographical characteristics and user activity levels of the regions, summarize the development trends of the user behaviors in each region, and output the results of the regional behavior trend analysis. Call the user behavior data of different regions, and compare the scoring trends of different regions. First, extract the scoring sequences of users in all regions, arrange them in chronological order, ensure that the data of all regions cover the same time period, and perform standardization processing on the scoring data to eliminate the differences in scoring scales among different regions. Then, analyze the scoring trends of users in each region, calculate the scoring fluctuation range of regional users, and identify the regions with relatively stable or large fluctuations in scoring. Next, calculate the similarity of the scoring patterns within each region, compare the deviation degrees of the scoring patterns in different regions, and evaluate the stability of the scoring patterns among regions. Subsequently, summarize the characteristics of the scoring trends of users in each region, match the geographical characteristics and user activity levels of each region, analyze the changing trends of the regional scoring patterns, and identify groups of regions with similar scoring patterns. Finally, organize and output the results of the regional behavior trend analysis for subsequent analysis and prediction.

[0042] Please refer to Figure 8 , the steps to obtain the evaluation results of the regional cultural dissemination effect are as follows: Based on the results of the regional behavior trend analysis, extract the scoring data of all users in the target region, filter the scoring records of the target topic, arrange them in chronological order, calculate the regional average score, score variance, skewness, and kurtosis of the score distribution of the topic, and summarize and obtain the acceptance data of the users in the target region for the topic; Based on the results of the regional behavior trend analysis, first filter the topics related to the target region, extract the interaction data of users in this region, filter out the browsing, liking, commenting, and sharing behaviors related to the topic, calculate the total interaction volume of each user on different topics, and classify them into different user type classifications to generate topic interaction data; Subsequently, according to the topic interaction data, count the user coverage and participation of each topic, calculate the attention change of the topic in different time periods, combine the preference information of users, standardize the interaction behaviors of different topics, and eliminate outliers to obtain the standardized topic interaction degree; Then, according to the standardized topic interaction degree, analyze the popularity distribution of the topic in the region, calculate the attention change in different time windows, further distinguish short-term popular topics and long-term continuously concerned topics, establish topic popularity time series data, and ensure that the dynamic change trend of the topic popularity over time is recorded; Finally, combine the topic popularity time series data, measure the acceptance of each topic in the target region, analyze the overlap degree of the audiences of different categories of topics, and judge the characteristics of topics with high acceptance.

[0043] Based on the acceptance data of the users in the target region for the topic, count the scoring changes in different time periods in the target region, calculate the growth rate and decay rate of the topic score, and analyze the change range of the score over time to evaluate the user penetration of the content dissemination, summarize the dissemination effect of the topic in different time ranges, and obtain the evaluation results of the content dissemination effect; First, extract the dissemination chains of each topic, analyze the user sharing paths, calculate the time intervals for topics to spread from the initial release to different social levels, count the secondary sharing rates of users during the topic dissemination process, and generate topic dissemination path data; Subsequently, based on the topic dissemination path data, calculate the penetration of the content among different user groups, evaluate the hierarchical diffusion of the content in the social network, count the dissemination depth and breadth, distinguish topics with high and low dissemination efficiency, eliminate the influence of local hotspots, and establish a content dissemination hierarchical model; Then, for the content dissemination hierarchical model, analyze the user attribute characteristics between different dissemination levels, calculate the influence of different user groups in the dissemination chain, evaluate the contribution of highly influential users to the topic dissemination effect, and obtain user dissemination influence indicators; Finally, combine the user dissemination influence indicators to evaluate the actual effect of content dissemination, identify the key nodes and bottlenecks in content dissemination, measure the actual reach rate of the content among different groups, and establish the evaluation results of content dissemination effect.

[0044] Based on the evaluation results of content dissemination effect, compare the dissemination fitness of different topics in the target area, combine the cultural characteristics of the target area, analyze the preferences of regional users for different topics, determine the cultural adaptability and summarize the cultural adaptation characteristics of the target area, and organize and output the evaluation results of regional cultural dissemination effect; First, collect historical data related to culture in the region, including language styles, value orientations, interest hotspots, etc., extract the language characteristics and expression methods of users during the content dissemination process, and conduct a matching analysis with the characteristics of regional culture to generate regional cultural expression characteristic data; Subsequently, based on the regional cultural expression characteristic data, calculate the matching degree between different topic contents and regional culture, count the acceptance of different cultural elements by users, analyze the adaptability of the content in different cultural backgrounds, and construct a cultural adaptability evaluation framework; Then, combine the cultural adaptability evaluation framework to quantify the cultural fitness of different topics, calculate the changing trend of regional cultural acceptance, and analyze whether the topics with long-term dissemination conform to the laws of regional cultural evolution to obtain the evaluation results of regional cultural fitness; Finally, according to the evaluation results of regional cultural fitness, combine the evaluation results of content dissemination effect, analyze the long-term impact of different content types in regional dissemination, measure the sustainability of cultural dissemination, and generate the evaluation results of regional cultural dissemination effect.

[0045] The above are only the preferred embodiments of the present invention, and do not limit the present invention in other forms. Any person skilled in the art may use the technical content disclosed above to make changes or modifications into equivalent embodiments with equivalent changes and apply them to other fields. However, as long as the technical solution content of the present invention is not departed from, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention still belong to the protection scope of the technical solution of the present invention.

Claims

1. Based on the regional cultural communication feasibility assessment system, it is characterized by: The system comprises: The data integration and formatting module extracts user video click records, viewing time and rating data, organizes and enters the video click records, viewing time and rating data into the database, creates data rows for each user and topic, verifies the integrity and accuracy of the data, and generates user activity feature data; The similarity matrix generation module counts the user rating patterns according to the user activity feature data, performs normalization processing on the rating data of each user, calculates the correlation coefficient, compares the rating vectors between different users one by one according to the calculation results, analyzes the similarity of the ratings of all users on the corresponding topics, and constructs and generates a similar user analysis matrix; The prediction analysis module analyzes the potential popularity of each unrated topic based on the similar user analysis matrix, predicts the rating of each topic, integrates user preference patterns, and outputs a predicted topic acceptance index; The behavior data analysis module groups the user behavior data according to the predicted topic acceptance index, performs trend analysis on each group data, identifies common and abnormal behavior patterns, reveals the behavior characteristics of each group according to the identification results, and corresponds them with the region to obtain the regional behavior trend analysis results; The communication impact assessment module measures the acceptance of the topic in the target area based on the regional behavior trend analysis results, evaluates the actual effect of content dissemination, determines the regional cultural adaptability, and generates regional cultural communication effect evaluation results.

2. The regional cultural communication feasibility assessment system according to claim 1 is characterized in that: The steps for acquiring the user activity feature data are as follows: Collect user click records, viewing time and rating data, filter all viewing behavior data of the same user in the same time window, sort by user ID, topic ID and timestamp, and use the formula: ； Calculate the user's effective viewing time , generate user effective viewing time data, where, Representing the user The duration of viewing behavior, Representative The scoring weight of each viewing behavior, Represents the total number of viewing behaviors; Based on the user's effective viewing time data, match the user's click behavior and rating information, establish a user's viewing behavior record within the target time window, classify and organize the user's viewing time, click count and rating results, and construct a user viewing behavior data row according to the user ID and topic ID; Based on the user viewing behavior data rows, screen abnormal data and erroneous data, match existing user behavior patterns, verify data integrity, create corresponding user activity feature data for each user and topic, and enter the data into the database.

3. The regional cultural communication feasibility assessment system according to claim 2 is characterized in that: The calculation steps of the correlation coefficient are: Based on the user activity feature data, extract the rating records of each user, collect users with complete rating data, calculate the average rating and standard deviation of each user, and generate user rating pattern data; According to the user rating mode data, the minimum and maximum values ​​of the rating data of each user are collected, the user rating data is mapped to a standard range, and a normalized rating value is calculated, and the user normalized rating data is sorted and output; Combined with the user normalized rating data, based on the covariance and standard deviation of the rating vector, the formula is adopted: ； Calculate the correlation coefficient between user rating vectors , get the user rating correlation coefficient data, where, Representative User and users The correlation coefficient between the scores, Representative User In the The normalized score value on the score items, Representative User In the The normalized score value on the score items, Representative User The mean score of Representative User The mean score of Represents the total number of rated items.

4. The regional cultural communication feasibility assessment system according to claim 3 is characterized in that: The steps for obtaining the similar user analysis matrix are: According to the user rating correlation coefficient data, combined with the preset threshold, the user pairs with correlation coefficients greater than the preset threshold are screened, and the formula is used: ； Counting users and users The rating vector distance between , get the user rating vector similarity data, where, Representative User In the The normalized score value on the score items, Representative User In the The normalized score value on the score items, Represents the total number of scoring items; According to the user rating vector similarity data, analyzing the consistency of all users' ratings on each topic, comparing all users' rating vectors one by one, analyzing the overlap of the rating vectors on a specific topic, and obtaining users' rating similarity data on the topic; Based on the similarity data of the user's ratings on the topic, a similar user analysis matrix is ​​constructed and outputted with the user identifier as the row index, the topic identifier as the column index, and the user's rating similarity on the topic as the filling matrix element.

5. The regional cultural communication feasibility assessment system according to claim 4 is characterized in that: The steps for obtaining the predicted topic acceptance index are: Based on the similar user analysis matrix, the topics that have not been rated by each user are screened, the rating data of similar users on the corresponding topics are extracted, the rating bias is identified in combination with the rating behaviors of similar users, and the potential popularity data of the topics are established; Based on the potential popularity data of the topic, a score prediction is made for each topic, the changes in the interest of the target user in different topics are analyzed, the historical score records and topic preferences of the target user are compared, and the topic score prediction data is generated in combination with the fluctuation analysis of the user's score on the topic; Based on the topic rating prediction data, the rating tendencies of different users on each topic are analyzed, the rating characteristics of different user groups are classified, the user interest distribution is determined, the overall rating trend is summarized, and the predicted topic acceptance index is output.

6. The regional cultural communication feasibility assessment system according to claim 5 is characterized in that: The steps of identifying the behavior pattern are: Based on the predicted topic acceptance index, the user's rating data on different topics is collected, and the users are classified according to their rating characteristics, rating stability and rating range, and the average and rating change range of the different user ratings are calculated. The users are grouped according to the similarity of the rating patterns, and the user behavior grouping data is output; Based on the user behavior grouping data, the rating trend of each group is analyzed, and the average rating value of the group at consecutive time points is extracted using the formula: ； Calculate the group in the time period arrive The rate of change of rating within , obtain group behavior trend data, where, Representative Group In time The average rating of Representative Group In time The average rating of represents a time interval; Based on the group behavior trend data, the scoring patterns within the group are compared, common behavior patterns with stable score changes are screened, individuals with abnormal score changes are extracted, the scoring behavior of abnormal users is analyzed, the score deviation characteristics are summarized, and the common and abnormal behavior pattern data are sorted and output.

7. The regional cultural communication feasibility assessment system according to claim 6 is characterized in that: The steps for obtaining the regional behavior trend analysis results are as follows: Based on the common and abnormal behavior pattern data, analyze the rating trend characteristics of each group, extract the stability of the rating changes within the group, the rating deviation and the rating fluctuation range, analyze the consistency of user behavior within the group, compare the rating preference distribution of users within the group, and summarize and generate group behavior feature data; Based on the group behavior feature data, the geographical area information of the group users is matched, the behavior patterns of the user groups in different regions are extracted, the rating trends of the users in the regions are compared, the rate of change of the regional user ratings is counted, the stability of the user behavior features within the regions is analyzed, the rating differences between different regions are evaluated, and the regional user behavior data is obtained; Combined with the regional user behavior data, the behavioral characteristics of the different regions are analyzed, and the rating preferences of the regional user groups are summarized, the geographical characteristics and user activity of the regions are matched, the user behavior development trends of each region are summarized, and the regional behavior trend analysis results are output.

8. The regional cultural communication feasibility assessment system according to claim 7 is characterized in that: The steps for obtaining the regional cultural communication effect evaluation results are as follows: Based on the regional behavior trend analysis results, extract the rating data of all users in the target area, filter the rating records of the target topic, and arrange them in chronological order, calculate the regional average rating, rating variance, and skewness and kurtosis of the rating distribution of the topic, and summarize and obtain the acceptance data of the topic by users in the target area; Based on the target area user acceptance data of the topic, statistics are collected on the score changes in different time periods in the target area, the growth rate and decay rate of the topic score are calculated, and the change range of the score over time is analyzed to evaluate the user penetration of content dissemination, summarize the dissemination effect of the topic within different time ranges, and obtain the content dissemination effect evaluation results; Based on the content communication effect evaluation results, the communication adaptability of different topics in the target area is compared, and combined with the cultural characteristics of the target area, the preferences of regional users for different topics are analyzed, the cultural adaptability is determined and the cultural adaptation characteristics of the target area are summarized, and the regional cultural communication effect evaluation results are sorted and output.

Images