Marketing video management system based on data analysis

Abstract

The invention relates to the technical field of marketing video management, and particularly discloses a marketing video management system based on data analysis, and the system comprises an acquisition module which is used for extracting music clips, text contents, portrait images and non-portrait images in a marketing video; the infringement risk analysis module is used for sequentially comparing and analyzing the acquired music clips, text contents, portrait images and non-portrait images with corresponding copyright libraries; performing comprehensive analysis to obtain a comprehensive infringement risk value corresponding to the marketing video; according to the method, the marketing videos with infringement risks are further subjected to hierarchical processing, the infringement screening efficiency and accuracy are improved, the dynamic risk threshold value is controlled and adjusted according to the historical infringement data of the publisher, a whole-process closed loop from content collection, risk analysis, response disposal to threshold value optimization is achieved, and the method is suitable for popularization and application. The infringement detection is comprehensive, the response strategy is flexible, the resource distribution is reasonable, and the problem that the marketing video infringement supervision is difficult is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of marketing video management, and specifically relates to a marketing video management system based on data analysis. Background Art

[0003] In the existing marketing video management process, there are many deficiencies in dealing with infringement supervision. For example, when faced with a large number of videos, the manual review method not only takes time and effort, but also is prone to missed judgments and misjudgments. And the system review method can only detect infringement of a certain type of content (such as text or images) in the video, unable to achieve multi-dimensional comprehensive analysis, and is also prone to missed judgments and misjudgments. Moreover, the infringement risk assessment threshold of the existing system is fixed and cannot be dynamically adjusted according to the historical compliance situation of the video publisher, resulting in insufficient supervision intensity and overly strict review of new users. Summary of the Invention

[0004] The purpose of the present invention is to provide a marketing video management system based on data analysis, and solve the following technical problems:

[0005] How to conduct multi-dimensional copyright comparison and analysis on marketing videos, combine user feedback confidence and data related to historical marketing video infringement of the publisher for dynamic risk assessment, and solve the problem of difficult infringement supervision of marketing videos.

[0006] The purpose of the present invention can be achieved through the following technical solutions:

[0007] A marketing video management system based on data analysis, the system includes:

[0008] A collection module, which extracts music segments, text content, portrait images and non-portrait images from marketing videos;

[0009] An infringement risk analysis module, which compares and analyzes the collected music segments, text content, portrait images and non-portrait images with the corresponding copyright libraries in sequence; and comprehensively analyzes to obtain the comprehensive infringement risk value corresponding to the marketing video;

[0010] An artificial feedback module, which collects and analyzes relevant user negative feedback data of marketing videos, and generates the user feedback confidence corresponding to the marketing video;

[0011] The risk response module compares the comprehensive infringement risk value with the preset infringement risk threshold range. When the comprehensive infringement risk value is greater than the upper limit of the preset infringement risk threshold range, it is determined that the marketing video has an infringement risk; when the comprehensive infringement risk value is within the preset infringement risk threshold range, a multi-dimensional infringement risk value is generated based on the user feedback confidence level of the marketing video. If the multi-dimensional infringement risk value is greater than the dynamic risk threshold, the corresponding marketing video is assigned to the manual review pool for manual review of its infringement risk; the marketing video determined to have an infringement risk is taken off the shelf;

[0012] The dynamic adjustment module adjusts the dynamic risk threshold according to the historical marketing video infringement-related data of the marketing video publisher.

[0013] Further, the process of comparing and analyzing the music segment with the corresponding copyright library includes:

[0014] Extract the audio features of the music segment in the marketing video, where the audio features include melody features, rhythm features, harmony features, and timbre features;

[0015] Convert the audio features into audio feature vectors;

[0016] Calculate the first cosine similarity between the audio feature vector and the audio feature vectors of each music work in the copyright library;

[0017] If there is any first cosine similarity greater than the preset music similarity threshold, it is determined that the marketing video has an infringement risk.

[0018] Further, the process of comparing and analyzing the text content with the corresponding copyright library includes:

[0019] Perform word segmentation on the text content in the marketing video and extract keywords;

[0020] Calculate the text similarity between the text content and each text work in the copyright library, and the text similarity calculation uses the TF-IDF algorithm combined with the cosine similarity formula;

[0021] If there is any text similarity greater than the preset text similarity threshold, it is determined that the marketing video has an infringement risk.

[0022] Further, the process of comparing and analyzing the portrait image with the corresponding copyright library includes:

[0023] Extract the facial feature points of the portrait image;

[0024] Convert the facial feature points into facial feature vectors;

[0025] Calculate the Euclidean distance between the facial feature vector and the feature vectors of each portrait image in the copyright library;

[0026] If there is any Euclidean distance less than the preset portrait similarity threshold, it is determined that the marketing video has an infringement risk.

[0027] Further, the process of comparing and analyzing the non-portrait image with the corresponding copyright library includes:

[0028] Extract the visual features of the non-portrait image, and the visual features include color features, texture features, and shape features;

[0029] Fuse the visual features into a visual feature vector;

[0030] Calculate the second cosine similarity between the visual feature vector and the feature vectors of each non-portrait image in the copyright library;

[0031] If there is any second cosine similarity greater than the preset image similarity threshold, it is determined that the marketing video has an infringement risk.

[0032] Further, the process of obtaining the comprehensive infringement risk value includes:

[0033] Perform a weighted sum of the infringement risk values of the music segment, the infringement risk value of the text content, the infringement risk value of the portrait image, and the infringement risk value of the non-portrait image to obtain the infringement risk value;

[0034] The calculation formula of the infringement risk value is as follows:

[0035]

[0036]

[0037]

[0038]

[0039]

[0040] Among them, is the comprehensive infringement risk value of the marketing video, is the infringement risk value of the music segment, is the infringement risk value of the text content, is the infringement risk value of the portrait image, is the infringement risk value of the non-portrait image, is the first cosine similarity between the music segment and the i-th music work in the copyright library, is the music infringement risk weight coefficient, is the text similarity between the text content and the j-th text work in the copyright library, is the text infringement risk weight coefficient, is the Euclidean distance between the portrait image and the k-th portrait image in the copyright library, is the portrait image infringement risk weight coefficient, is the second cosine similarity between the non-portrait image and the l-th non-portrait image in the copyright library, is the non-portrait image infringement risk weight coefficient.

[0041] Furthermore, the process of obtaining the multi-dimensional infringement risk value includes:

[0042] Extract user negative feedback data related to the marketing video, and the user negative feedback data includes negative comments on marketing video infringement and infringement negative bullet screens;

[0043]

[0044] Analyze and calculate the user feedback confidence level corresponding to the marketing video through formula (6) ;

[0045] where N is the number of items of user negative feedback data, , is the quantity of the p-th user negative feedback, is the weight coefficient of the p-th negative feedback;

[0046]

[0047] Analyze and calculate the multi-dimensional infringement risk value through formula (7) .

[0048] Furthermore, the process of adjusting the preset dynamic risk threshold includes:

[0049] Obtain the historical marketing video infringement-related data of the marketing video publisher, and the historical marketing video infringement-related data includes the total number of historical published marketing videos, the number of historical marketing videos with infringement risks, and the number of historical confirmed infringement marketing videos;

[0050]

[0051] Analyze and calculate the compliance index of the publisher's published marketing videos through formula (8) ;

[0052] where M is the total number of historical published marketing videos, is the number of historical marketing videos with infringement risks, is the comprehensive infringement risk value of the u-th marketing video with infringement risks;

[0053] Adjust the preset dynamic threshold according to the compliance index S, and the adjustment formula is:

[0054]

[0055] Wherein, is the dynamic risk threshold, is the preset benchmark risk threshold.

[0056] Advantages of the present invention:

[0057] (1) In the present invention, the acquisition module extracts multi-dimensional data from the marketing video, the infringement risk analysis module analyzes whether the multi-dimensional data is infringed, and the risk analysis module collects the hierarchical processing logic to perform hierarchical response processing on the infringement risk degree of the marketing video. During the processing, the manual feedback module generates feedback confidence according to the negative feedback data of the marketing video, and further performs hierarchical processing on the marketing video with medium infringement risk, improving the efficiency and accuracy of infringement screening, and controlling and adjusting the dynamic risk threshold according to the historical infringement data of the publisher, realizing a full-process closed loop from content acquisition, risk analysis, response handling to threshold optimization, making the infringement detection comprehensive, the response strategy flexible and the resource allocation reasonable, and solving the problem of difficult infringement supervision of marketing videos. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] The present invention will be further described below with reference to the accompanying drawings.

[0059] Figure 1 is a schematic block diagram of a marketing video management system based on data analysis proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0060] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0061] Please refer to Figure 1 as shown. In one embodiment, a marketing video management system based on data analysis is provided, and the system includes:

[0062] An acquisition module, through audio-visual parsing technology, separates and extracts music segments (such as BGM, sound effects), text content (such as subtitles, slogans), portrait images (images including human faces), and non-portrait images (such as scene maps, icons, illustrations, etc.) from the marketing video;

[0063] The infringement risk analysis module compares and analyzes the collected music segments, text content, portrait images, and non-portrait images with the corresponding copyright libraries (such as music copyright libraries, text work libraries, portrait copyright libraries, and image material libraries) in sequence; and obtains the comprehensive infringement risk value corresponding to the marketing video through comprehensive analysis using a multi-dimensional feature matching algorithm.

[0064] The manual feedback module collects negative feedback data (such as negative comments, etc.) from users on the marketing video, quantifies the feedback intensity of users on infringement issues through natural language processing (NLP) and data statistical analysis, and generates a feedback confidence level.

[0065] The risk response module adopts a hierarchical processing logic to compare the comprehensive infringement risk value with the preset infringement risk threshold range. When the comprehensive infringement risk value is greater than the upper limit of the preset infringement risk threshold range, it indicates that the marketing video is a high-infringement-risk marketing video, and it is determined that the marketing video has an infringement risk; when the comprehensive infringement risk value is within the preset infringement risk threshold range, it indicates that the marketing video is a medium-infringement-risk marketing video, and then a multi-dimensional infringement risk value is generated based on the user feedback confidence level of the marketing video. If the multi-dimensional infringement risk value is greater than the dynamic risk threshold, the corresponding marketing video is assigned to the manual review pool for manual review of its infringement risk, dynamically determining whether to enter the manual review pool to improve the efficiency and accuracy of infringement screening; the marketing videos determined to have an infringement risk are taken off the shelves.

[0066] The dynamic adjustment module dynamically adjusts the risk threshold according to the historical marketing video infringement-related data (such as infringement frequency, severity) of the marketing video publisher through an algorithm, so that the system adapts to the compliance levels of different publishers and improves the scientificity of threshold setting.

[0067] Through the above technical solutions, this embodiment provides a marketing video management system based on data analysis. The system extracts multi-dimensional data from the marketing video through the acquisition module, analyzes whether the multi-dimensional data is infringing through the infringement risk analysis module, and performs hierarchical response processing on the infringement risk degree of the marketing video through the risk analysis module's hierarchical processing logic. During the processing, the manual feedback module generates a feedback confidence level based on the negative feedback data of the marketing video, further performs hierarchical processing on the medium-infringement-risk marketing videos, improves the efficiency and accuracy of infringement screening, and controls and adjusts the dynamic risk threshold according to the publisher's historical infringement data, realizing a full-process closed-loop from content acquisition, risk analysis, response handling to threshold optimization, making the infringement detection comprehensive, the response strategy flexible, and the resource allocation reasonable, and solving the problem of difficult supervision of marketing video infringement.

[0068] In one embodiment, the process of comparing and analyzing the music segment with the corresponding copyright library includes:

[0069] Extract the audio features of the music segment in the marketing video. Using digital signal processing technology, separate and quantify the melody features (pitch sequence), rhythm features (beat, tempo), harmony features (chord progression), and timbre features (instrument type, timbre texture) from the music segment to form a multi-dimensional feature vector.

[0070] Calculate the first cosine similarity between the audio feature vector and the audio feature vectors of each music work in the copyright library. The closer the value of the first cosine similarity is to 1, the more similar the audio structures are.

[0071] If there is any first cosine similarity greater than the preset music similarity threshold, and the preset music similarity threshold can be obtained by presetting according to experience, it is determined that the marketing video has an infringement risk, and then the marketing video is taken off the shelves.

[0072] The process of comparing and analyzing the text content with the corresponding copyright library includes:

[0073] Perform word segmentation on the text content in the marketing video. Use NLP technology to perform Chinese word segmentation on the text content (such as subtitles, voiceovers), and extract the core keywords (such as brand names, creative concept words) after removing stop words.

[0074] Calculate the text similarity between the text content and each text work in the copyright library. The text similarity calculation uses the TF-IDF algorithm combined with the cosine similarity formula. The specific process is as follows: Calculate the term frequency (TF) of the keywords in the text and the inverse document frequency (IDF) in the copyright library to highlight the distinctive words. After converting the text into a TF-IDF vector, calculate the cosine value of the vector angle with the copyright library text. The larger the value, the more similar the text structures are.

[0075] If there is any text similarity greater than the preset text similarity threshold, and the preset text similarity threshold can be obtained by presetting according to experience, it is determined that the marketing video has an infringement risk, and then the marketing video is taken off the shelves.

[0076] The process of comparing and analyzing the portrait image with the corresponding copyright library includes:

[0077] Use computer vision technology (such as OpenCV, FaceNet) to detect the facial key points (such as 68 landmark points like the corners of the eyes, the tip of the nose, the corners of the mouth, etc.) in the portrait image and extract the facial feature points of the portrait image.

[0078] Convert the facial feature points into a facial feature vector.

[0079] Calculate the Euclidean distance between the facial feature vector and the feature vectors of each portrait image in the copyright library. The smaller the distance, the higher the facial similarity.

[0080] If there is any Euclidean distance less than the preset portrait similarity threshold, which can be obtained by presetting according to experience, it is determined that the marketing video has an infringement risk, and then the marketing video is taken off the shelf.

[0081] The process of comparing and analyzing the non-portrait image with the corresponding copyright library includes:

[0082] Extract the visual features of the non-portrait image. The visual features include color features, texture features, and shape features. Color features: Quantify the color composition of the image through color histograms and dominant color distributions; Texture features: Use the gray-level co-occurrence matrix (GLCM) to extract features such as image texture roughness and directionality; Shape features: Extract geometric shape parameters (such as circularity and aspect ratio) through contour detection and Hough transform.

[0083] Fuse the visual features into a visual feature vector.

[0084] Calculate the second cosine similarity between the visual feature vector and the feature vectors of each non-portrait image in the copyright library. The closer the value of the second cosine similarity is to 1, the more similar the non-portrait images are.

[0085] If there is any second cosine similarity greater than the preset image similarity threshold, which can be obtained by presetting according to experience, it is determined that the marketing video has an infringement risk, and then the marketing video is taken off the shelf.

[0086] The process of obtaining the comprehensive infringement risk value includes:

[0087] Weighted sum of the infringement risk value of the music segment, the infringement risk value of the text content, the infringement risk value of the portrait image, and the infringement risk value of the non-portrait image to obtain the infringement risk value.

[0088] The calculation formula of the infringement risk value is as follows:

[0089]

[0090]

[0091]

[0092]

[0093]

[0094] Among them, is the comprehensive infringement risk value of the marketing video, is the infringement risk value of the music segment, is the infringement risk value of the text content, is the infringement risk value of the portrait image, is the infringement risk value of the non - portrait image, is the first cosine similarity between the music segment and the i - th music work in the copyright library, is the music infringement risk weight coefficient, is the text similarity between the text content and the j - th text work in the copyright library, is the text infringement risk weight coefficient, is the Euclidean distance between the portrait image and the k - th portrait image in the copyright library, is the portrait image infringement risk weight coefficient, is the second cosine similarity between the non - portrait image and the l - th non - portrait image in the copyright library, is the non - portrait image infringement risk weight coefficient, , , , reflect the infringement sensitivity of different content types, which can be obtained by presetting according to experience.

[0095] Through the above technical solution, this embodiment provides a method for quantitatively evaluating infringement risk based on multi - modal feature extraction and weighted fusion. The method constructs a unified comprehensive infringement risk assessment system by weighted summing the infringement risk values of music segments, text content, portrait images, and non - portrait images, realizes the full - element infringement risk scanning of marketing videos, can accurately quantify the infringement risk of marketing videos, and is convenient for subsequent hierarchical disposal mechanisms (direct removal, manual review, temporary non - processing) based on the comprehensive risk value, so as to achieve optimal resource allocation.

[0096] In one embodiment, the process of obtaining the multi - dimensional infringement risk value includes:

[0097] Extract the user negative feedback data related to the marketing video to form a multi - source feedback data set. The user negative feedback data includes negative comments on marketing video infringement and negative bullet screens. Comments or bullet screens that mention keywords such as "infringement" and "plagiarism" in the comments or bullet screens are negative comments on marketing video infringement and negative bullet screens;

[0098]

[0099] Analyze and calculate the user feedback confidence corresponding to the marketing video through formula (6) ;

[0100] where N is the number of items of user negative feedback data, , is the quantity of the p - th user negative feedback, which is obtained by counting the user negative feedback data. is the weight coefficient of the p-th negative feedback, reflecting the credibility of user feedback, which can be obtained by presetting according to experience;

[0101]

[0102] The multi-dimensional infringement risk value is obtained through analysis and calculation by formula (7) .

[0103] Through the above technical solution, this embodiment provides a multi-dimensional risk assessment method that combines system automation analysis and user supervision. By collecting and quantifying user negative feedback, the subjective evaluation is converted into an objective confidence index, which is cross-validated with the results of automated infringement detection, improving the reliability of the infringement risk assessment results.

[0104] In one embodiment, the process of adjusting the preset dynamic risk threshold includes:

[0105] Obtain the historical marketing video infringement-related data of the marketing video publisher, where the historical marketing video infringement-related data includes the total number of historical published marketing videos, the number of historical marketing videos with infringement risks, and the number of historical confirmed infringing marketing videos;

[0106]

[0107] The compliance index of the publisher's published marketing videos is obtained through analysis and calculation by formula (8) ;

[0108] Among them, M is the total number of historical published marketing videos, which can be obtained by counting the marketing videos published by the publisher, is the number of historical marketing videos with infringement risks, which can be counted according to the historical records of infringement analysis, is the comprehensive infringement risk value of the u-th marketing video with infringement risks, which can be obtained according to the above infringement risk value;

[0109] Adjust the preset dynamic threshold according to the compliance index S, and the adjustment formula is:

[0110]

[0111] Among them, is the dynamic risk threshold, is the preset benchmark risk threshold, which can be obtained by presetting according to experience.

[0112] Through the above technical solution, this embodiment provides a dynamic threshold adjustment method based on the publisher's historical records. By quantitatively analyzing the publisher's infringement history, the adaptive adjustment of the risk threshold is realized, achieving the purpose of optimizing system resource allocation (prioritizing the review of high-risk subjects) and personalizing management strategies.

[0113] The above has described in detail an embodiment of the present invention, but the above content is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. Any equivalent changes and improvements made within the scope of the application of the present invention shall still fall within the scope covered by the patent of the present invention.

Claims

1. A marketing video management system based on data analysis, characterized in that, The system includes: A collection module that extracts music segments, text content, portrait images, and non-portrait images from marketing videos; An infringement risk analysis module that sequentially compares and analyzes the collected music segments, text content, portrait images, and non-portrait images with the corresponding copyright libraries; and comprehensively analyzes to obtain a comprehensive infringement risk value corresponding to the marketing video; An artificial feedback module that collects and analyzes relevant user negative feedback data of the marketing video to generate a user feedback confidence level corresponding to the marketing video; A risk response module that compares the comprehensive infringement risk value with a preset infringement risk threshold range. When the comprehensive infringement risk value is greater than the upper limit of the preset infringement risk threshold range, it is determined that the marketing video has an infringement risk; when the comprehensive infringement risk value is within the preset infringement risk threshold range, a multi-dimensional infringement risk value is generated based on the user feedback confidence level of the marketing video. If the multi-dimensional infringement risk value is greater than the dynamic risk threshold, the corresponding marketing video is assigned to an artificial review pool for manual review of its infringement risk; the marketing video determined to have an infringement risk is taken off the shelf; A dynamic adjustment module that adjusts the dynamic risk threshold according to the historical marketing video infringement-related data of the marketing video publisher.

2. The marketing video management system based on data analysis according to claim 1, wherein The process of comparing and analyzing the music segment with the corresponding copyright library includes: Extracting the audio features of the music segment in the marketing video, where the audio features include melody features, rhythm features, harmony features, and timbre features; Converting the audio features into audio feature vectors; Calculating the first cosine similarity between the audio feature vectors and the audio feature vectors of each music work in the copyright library; If there is any first cosine similarity greater than the preset music similarity threshold, it is determined that the marketing video has an infringement risk.

3. The marketing video management system based on data analysis according to claim 2, wherein The process of comparing and analyzing the text content with the corresponding copyright library includes: Performing word segmentation on the text content in the marketing video and extracting keywords; Calculating the text similarity between the text content and each text work in the copyright library, where the text similarity calculation uses the TF-IDF algorithm combined with the cosine similarity formula; If there is any text similarity greater than the preset text similarity threshold, it is determined that the marketing video has an infringement risk.

4. The marketing video management system based on data analysis according to claim 3, characterized in that, The process of comparing and analyzing the portrait image with the corresponding copyright library includes: Extracting the facial feature points of the portrait image; Converting the facial feature points into facial feature vectors; Calculating the Euclidean distance between the facial feature vectors and the feature vectors of each portrait image in the copyright library; If there is any Euclidean distance less than the preset portrait similarity threshold, it is determined that the marketing video has an infringement risk.

5. The marketing video management system based on data analysis according to claim 4, characterized in that, The process of comparing and analyzing the non-portrait image with the corresponding copyright library includes: Extracting the visual features of the non-portrait image, where the visual features include color features, texture features, and shape features; Fusing the visual features into visual feature vectors; Calculating the second cosine similarity between the visual feature vectors and the feature vectors of each non-portrait image in the copyright library; If there is any second cosine similarity greater than the preset image similarity threshold, it is determined that the marketing video has an infringement risk.

6. The marketing video management system based on data analysis according to claim 5, wherein The process of obtaining the comprehensive infringement risk value includes: The infringement risk value of a music segment, the infringement risk value of text content, the infringement risk value of a portrait image, and the infringement risk value of a non-portrait image are weighted and summed to obtain the infringement risk value; The calculation formula of the infringement risk value is as follows: Among them, is the comprehensive infringement risk value of the marketing video, is the infringement risk value of the music segment, is the infringement risk value of the text content, is the infringement risk value of the portrait image, is the infringement risk value of the non - portrait image, is the first cosine similarity between the music segment and the i - th music work in the copyright library, is the music infringement risk weight coefficient, is the text similarity between the text content and the j - th text work in the copyright library, is the text infringement risk weight coefficient, is the Euclidean distance between the portrait image and the k - th portrait image in the copyright library, is the portrait image infringement risk weight coefficient, is the second cosine similarity between the non - portrait image and the l - th non - portrait image in the copyright library, is the non - portrait image infringement risk weight coefficient.

7. The marketing video management system based on data analysis according to claim 6, characterized in that The process of obtaining the multi-dimensional infringement risk value includes: Extract user negative feedback data related to the marketing video, where the user negative feedback data includes negative comments on marketing video infringement and infringement negative bullet screens; The confidence level of user feedback corresponding to the marketing video is obtained through analytical calculation using formula (6). ; where N is the number of items of user negative feedback data, , is the quantity of the p-th item of user negative feedback, is the weight coefficient of the p-th item of negative feedback; The multi-dimensional infringement risk value is obtained through analysis and calculation by formula (7). .

8. The marketing video management system based on data analysis according to claim 7, characterized in that The process of adjusting the preset dynamic risk threshold includes: Obtain historical marketing video infringement-related data of the marketing video publisher, where the historical marketing video infringement-related data includes the total number of historical published marketing videos, the number of historical marketing videos with infringement risks, and the number of historical confirmed infringement marketing videos; The compliance index of the marketer's released marketing video is obtained through analysis and calculation by formula (8). ; where M is the total number of historical released marketing videos, is the number of historical marketing videos with infringement risks, is the comprehensive infringement risk value of the u-th marketing video with infringement risks; Adjust the preset dynamic threshold according to the compliance index S, and the adjustment formula is: Among them, is the dynamic risk threshold, is the preset reference risk threshold.

Images