A satisfaction evaluation method and system based on network data crawling

Abstract

The present invention relates to the field of electronic digital data processing technology, specifically to a satisfaction evaluation method and system based on network data capture, the method comprising: capturing network evaluation data of a target product, including image feedback data and user evaluation data; analyzing the network evaluation data of the target product to obtain a product visual quality evaluation index and a user feedback quality evaluation index at each time monitoring point; comprehensively analyzing the product visual quality evaluation index and the user feedback quality evaluation index at each time monitoring point to obtain product satisfaction evaluation index parameters at each time monitoring point, analyzing user satisfaction, and providing feedback based on the analysis results. The present invention provides a satisfaction evaluation method and system based on network data capture, and network data capture is based on the real behavior and feedback of users on the Internet, which makes the evaluation results more objective and improves the accuracy of the evaluation results.

Description

Technical Field

[0001] The present invention relates to the technical field of electronic digital data processing, and in particular to a satisfaction evaluation method and system based on network data capture. Background Art

[0002] With the development of social economy and the improvement of people's living standards, consumers have higher and higher requirements for the quality of goods and services, and service quality has become a reflection of the core competitiveness of enterprises. Satisfaction evaluation methods can help enterprises fully and deeply understand service quality, discover and solve service quality problems, and improve service levels.

[0003] For example, the invention patent with announcement number: CN111475703B discloses an analysis method for capturing specific network data. The analysis method for capturing specific network data is that the data first enters the algorithm system through filtering conditions, and the data is first processed using the feature code model. If the target data is not found, it is transferred to the regional model for further processing. If it still fails, it is deemed that the target data does not exist. Compared with the prior art, this application has stronger data targeting and can capture data in a data-closed environment, which is conducive to developers quickly implanting applications into the system and eliminating the data docking process. The present invention simultaneously enables two models to analyze and extract data. The advantages and disadvantages of the two models complement each other, and reliable capture of data with certain characteristics is achieved.

[0004] For example, the invention patent with announcement number: CN106326447B discloses a method and system for detecting data captured by a crowdsourced web crawler, which can ensure that the data captured by the crowdsourced web crawler is authentic and reliable. This method for detecting data captured by a crowdsourced web crawler uses the server as the inspection center for the crawler client's crawling results. The crawler client uploads the captured page content to the inspection center, and the inspection center compares the content uploaded by multiple crawler clients. If the results are the same, credit points are added to each crawler client; if the results are different, the task is issued again, and the crawler clients are re-inspected to distinguish the pros and cons, and then the corresponding credit points are added or subtracted; the credit points indicate the reliability of the crawler client, and crawler clients with high credit points are preferentially selected to complete the crawling task. A detection system for crowdsourced web crawler crawling data is also provided.

[0005] Based on the above solution, it is found that there are still some shortcomings in the satisfaction evaluation. Specifically, the data obtained in the satisfaction data analysis method of the target products on the Internet is too single. The single-dimensional network crawled data is easy to be tampered with, which may distort the satisfaction evaluation results and cause distorted evaluation results. Summary of the invention

[0006] In view of the deficiencies in the prior art, the present invention provides a satisfaction evaluation method and system based on network data crawling, which can effectively solve the problems involved in the above-mentioned background technology.

[0007] To achieve the above objectives, the present invention is implemented through the following technical solutions: The first aspect of the present invention provides a satisfaction evaluation method based on network data capture, including: capturing network evaluation data of the target product, wherein the network evaluation data of the target product includes image feedback data and user evaluation data of the target product.

[0008] The image feedback data of the target product is processed to obtain the product visual quality evaluation index at each time monitoring point.

[0009] The user evaluation data are classified, including user comments on each product at each time monitoring point, user ratings of each product, and page views of user comments on each product, and a user feedback quality evaluation index at each time monitoring point is obtained through comprehensive analysis.

[0010] A comprehensive analysis is conducted on the product visual quality evaluation index and user feedback quality evaluation index at each time monitoring point to obtain the product satisfaction evaluation index parameters at each time monitoring point. Based on the product satisfaction evaluation index parameters at each time monitoring point, user satisfaction is analyzed to obtain the analysis results and provide feedback.

[0011] As a further method, the image feedback data of the target product is processed to obtain the product visual quality evaluation index at each time monitoring point. The specific processing process is: deploy a number of time monitoring points, collect new product feedback images at each time monitoring point, overlap and compare each new product feedback image at each time monitoring point with the reference standard image in the product database, extract the pixel value of each pixel point of each new product feedback image at each time monitoring point and the corresponding pixel point of the reference standard image, and comprehensively analyze to obtain the product visual quality evaluation index at each time monitoring point.

[0012] As a further method, the user evaluation data is classified, and the specific process is: extracting each product user score from the user evaluation data, comparing each product user score with the reference critical score stored in the product database, if a product user score is greater than or equal to the reference critical score, then marking the product user comment corresponding to the product user score as a positive product comment, and counting the total number of product user comments and the number of positive product comments at each time monitoring point.

[0013] As a further method, the comprehensive analysis obtains the user feedback quality evaluation index at each time monitoring point, and the specific analysis process is: collecting the total page views of product user comments and the page views of product positive comments at each time monitoring point, performing a ratio operation on the total page views of product user comments and the page views of product positive comments at each time monitoring point, and marking the ratio result as a positive comment influencing factor;

[0014] Based on the user ratings of each product at each time monitoring point, the total number of product user reviews, the number of positive product reviews and the positive review influencing factors, a comprehensive analysis is conducted to obtain the user feedback quality evaluation index at each time monitoring point.

[0015] As a further method, the product satisfaction evaluation index parameters at each time monitoring point, the specific analysis process is: comprehensive analysis of the product visual quality evaluation index and the user feedback quality evaluation index at each time monitoring point to obtain product satisfaction evaluation index parameters, the product satisfaction evaluation index parameters are used to quantitatively evaluate user satisfaction with the target product.

[0016] As a further method, the user satisfaction is analyzed, the analysis results are obtained and feedback is provided, and the specific analysis process is: curve fitting is performed on the product satisfaction evaluation index parameters at each time monitoring point through polynomial regression to obtain the time series curve and time-dependent function of the product satisfaction evaluation index parameters at each time monitoring point, the most recent time monitoring point is marked as the current time monitoring point, the time-dependent function of the product satisfaction evaluation index parameters at the current time monitoring point is derived, the time derivative of the product satisfaction evaluation index parameters at the current time monitoring point is obtained, and it is marked as a reference index for the satisfaction change trend;

[0017] Compare the satisfaction change trend reference index with 0. If the satisfaction change trend reference index is greater than or equal to 0, the product user's satisfaction change trend is marked as a stable trend. If the satisfaction change trend reference index is less than 0, the product user's satisfaction change trend is marked as a downward trend.

[0018] The analysis results are fed back to relevant personnel, who then improve the product based on the actual content of user comments.

[0019] As a further method, the product visual quality evaluation index at each time monitoring point is a quantitative evaluation data obtained by comparing and analyzing the feedback images of each new product at each time monitoring point with the reference standard image, which is used to quantitatively evaluate the visual quality of the target product and provide a basis for evaluating the satisfaction of the target product.

[0020] As a further method, the product visual quality evaluation index at each time monitoring point is specifically expressed as:

[0021]

[0022] Where Y s represents the product visual quality evaluation index at the s-th time monitoring point, s represents the number of each time monitoring point, s = 1, 2, 3, ..., p, p represents the total number of time monitoring points, m represents the number of each newly added product feedback image, m = 1, 2, 3, ..., t, t represents the total number of newly added product feedback images, r represents the number of each pixel point, r = 1, 2, 3, ..., h, h represents the total number of pixels, X s→mr represents the pixel value of the rth pixel of the image of the mth target product at the sth time monitoring point, represents the pixel value of the rth pixel point of the reference standard image, ΔX represents the allowable deviation pixel value, and γ represents the set visual quality assessment correction factor.

[0023] As a further method, the product satisfaction evaluation index parameter at each time monitoring point is specifically expressed as follows:

[0024]

[0025] In the formula, represents the product satisfaction evaluation index parameter at the sth time monitoring point, softplus(x)=ln(1+e x ), this function is usually used to construct activation functions to introduce nonlinear transformations and help the model learn more complex features and patterns. It is widely used in various network architectures, such as convolutional neural networks, recurrent neural networks, and generative adversarial networks. Ys represents the product visual quality evaluation index at the s-th time monitoring point, Gs represents the user feedback quality evaluation index at the s-th time monitoring point, ε1 represents the satisfaction evaluation influence factor corresponding to the set visual quality evaluation index, and ε2 represents the satisfaction evaluation influence factor corresponding to the set user feedback quality evaluation index.

[0026] The second aspect of the present invention provides a satisfaction evaluation system based on network data capture, including: a network data acquisition module, an image feedback data processing module, a user evaluation data analysis module, a comprehensive analysis module and a product database.

[0027] Network data collection module, used to capture network evaluation data of target products;

[0028] An image feedback data processing module is used to process the image feedback data of the target product to obtain the product visual quality evaluation index at each time monitoring point;

[0029] The user evaluation data analysis module is used to obtain the user feedback quality evaluation index at each time monitoring point through comprehensive analysis based on user ratings, the number of user comments in each category, and the number of comment views;

[0030] Comprehensive analysis module, used to comprehensively analyze the product visual quality evaluation index and user feedback quality evaluation index at each time monitoring point, and obtain the satisfaction evaluation index parameters of the target product. According to the satisfaction evaluation index parameters of the target product, the target product is evaluated, the evaluation results are obtained and feedback is given;

[0031] The product database is used to store relevant evaluation data of the target product, including the reference standard image and the pixel values ​​of its corresponding pixel points, the allowable deviation pixel values, the reference critical score and the reference index of the satisfaction change trend.

[0032] Compared with the prior art, the embodiments of the present invention have at least the following advantages or beneficial effects:

[0033] (1) The present invention provides a satisfaction evaluation method and system based on network data capture. Network data capture is based on the real behavior and feedback of users on the Internet, which makes the evaluation results more objective. It avoids the subjective bias that may appear in traditional survey methods and improves the accuracy of the evaluation results. Through network data capture, a large amount of user feedback information can be quickly collected, which greatly improves the efficiency of satisfaction evaluation.

[0034] (2) The present invention analyzes the network feedback images of the target product and identifies the key features in the images, which is helpful for discovering the advantages and disadvantages in product design. Image processing can make the evaluation of user feedback more quantitative. Network feedback images can be cross-validated with other types of feedback. Multi-angle and multi-modal information integration helps to improve the reliability of the overall data. Combining network image feedback data with traditional text feedback data can provide more comprehensive and multi-dimensional satisfaction evaluation results.

[0035] (3) The present invention analyzes the user evaluation data of the target product. The user evaluation data covers a variety of usage scenarios and experiences, which helps to fully understand the performance of the product in actual use, including advantages and potential disadvantages. Analyzing the user evaluation data helps to improve the representativeness and practicality of the data. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] The present invention is further described using the accompanying drawings, but the embodiments in the accompanying drawings do not constitute any limitation to the present invention. A person skilled in the art can obtain other drawings based on the following drawings without creative work.

[0037] Figure 1 It is a schematic diagram of the method flow of the present invention.

[0038] Figure 2 It is a schematic diagram of system module connection of the present invention.

[0039] Figure 3 The figure is a schematic diagram of the parameter function relationship between the product visual quality evaluation index and the product satisfaction evaluation index involved in the embodiment of the present invention. DETAILED DESCRIPTION

[0040] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0041] Reference Figure 1 As shown, the first aspect of the present invention provides a satisfaction evaluation method based on network data capture, including: capturing network evaluation data of a target product, wherein the network evaluation data of the target product includes image feedback data and user evaluation data of the target product.

[0042] Specifically, the network evaluation data of the target product is crawled, including the image feedback data and user evaluation data of the target product. The image feedback data of the target product is obtained by crawling the target product using Python language and its rich library support (such as BeautifulSoup, Scrapy, Selenium, etc.), executing JavaScript to load the dynamic web page content of the target product, analyzing the web page structure, downloading the pictures in the web page using libraries such as requests, and extracting the URL of the pictures in the evaluation for processing. The user evaluation data is obtained by locating the user evaluation content through the Cascading Style Sheets selector or XPath expression and extracting the text information.

[0043] In a specific embodiment, the network evaluation data of the target product, especially the image feedback data and the user evaluation text, are captured and analyzed. The image feedback data and the user evaluation text provide a richer information dimension than the text. The image can directly show the actual status of the product, while the text evaluation provides detailed usage experience and opinions. Combining these two data types can obtain a more comprehensive view of product performance. Through text analysis, sentiment analysis can be performed to understand the user's satisfaction and emotional tendency with the product, so as to quickly identify positive and negative user feedback. Data processing can help identify trends in specific product features, such as which features are considered valuable, which problems occur frequently, and how these trends change over time. Through data mining technology, useful patterns and knowledge can be extracted from a large number of user reviews, such as user preferences for different aspects of the product. Natural language processing and image recognition technology can be used to automatically analyze user reviews, save human resources, and improve analysis efficiency. Personalized analysis can be performed according to different user groups or market segments to meet different business needs. Through data visualization tools, the analysis results can be intuitively displayed in the form of charts or dashboards to help decision makers understand the data more intuitively.

[0044] The image feedback data of the target product is processed to obtain the product visual quality evaluation index at each time monitoring point.

[0045] Specifically, the image feedback data of the target product is processed to obtain the product visual quality evaluation index at each time monitoring point. The specific processing process is: deploy a number of time monitoring points, collect new product feedback images at each time monitoring point, overlap and compare each new product feedback image at each time monitoring point with the reference standard image in the product database, extract the pixel value of each pixel point of each new product feedback image at each time monitoring point and the corresponding pixel point of the reference standard image, and comprehensively analyze to obtain the product visual quality evaluation index at each time monitoring point.

[0046] In a specific embodiment, the product visual quality evaluation index at each time monitoring point is specifically expressed as:

[0047]

[0048] Where Ys represents the product visual quality evaluation index at the s-th time monitoring point, s represents the number of each time monitoring point, s = 1, 2, 3, ..., p, p represents the total number of time monitoring points, m represents the number of each newly added product feedback image, m = 1, 2, 3, ..., t, t represents the total number of newly added product feedback images, r represents the number of each pixel point, r = 1, 2, 3, ..., h, h represents the total number of pixels, Xs → mr represents the pixel value of the r-th pixel point of the image of the m-th target product at the s-th time monitoring point, represents the pixel value of the rth pixel point of the reference standard image, ΔX represents the allowable deviation pixel value, and γ represents the set visual quality assessment correction factor.

[0049] Specifically, the product visual quality evaluation index at each time monitoring point is a quantitative evaluation data obtained by comprehensively analyzing the pixel values ​​of each pixel point of each new product feedback image at each time monitoring point and the corresponding pixel point of the reference standard image. It is used to quantitatively evaluate the visual quality of the target product and provide a basis for user satisfaction evaluation.

[0050] Furthermore, in this embodiment, the value range of the visual quality assessment correction factor is [0.1, 0.8]. A smaller visual quality assessment correction factor will make the formula less sensitive to the deviation of the pixel value. Even if the pixel deviation is large, the impact on the final evaluation index is relatively limited. This is suitable for scenarios with a high tolerance for changes in visual quality. A larger visual quality assessment correction factor will amplify the impact of pixel deviations, making the evaluation index react strongly to even small changes in pixel values. It is suitable for applications that have extremely high requirements for product quality and need to finely distinguish quality differences. The visual quality assessment correction factor is used to reduce calculation errors and improve accuracy.

[0051] It should be explained that, in addition to being calculated by the above formula, the product visual quality evaluation index at each time monitoring point in this embodiment can also be obtained by utilizing image processing software and algorithms, such as edge detection, filtering, segmentation, threshold processing, etc., to analyze the features in the image and identify defects, size, shape, color consistency, etc. in the product appearance. The system can also learn to identify the image features of high-quality and low-quality products and automatically evaluate the visual quality by training classifiers or neural network models (such as convolutional neural networks).

[0052] It should be explained that the newly added product feedback images at each time monitoring point in this embodiment are quality feedback images about the target product added at each time monitoring point, which are obtained by crawling network data through Python language and executing JavaScript to load dynamic web page content about the target product. All the newly added product feedback images in this embodiment are grayscale images, and grayscale images have only one channel. The pixel value is a scalar, which represents the brightness of the image. Then the pixel value of the corresponding pixel point of each newly added product feedback image is the brightness of the corresponding pixel point. By comparing pixel by pixel, any slight changes in the quality of the product image can be captured in detail, ensuring the comprehensiveness and meticulousness of the evaluation. Through statistical analysis of pixel values, abnormal areas in product images can be identified, which may represent product defects or damage. For example, if the pixel value of a certain area deviates significantly from the pixel distribution of a normal product image, it may indicate scratches, cracks or other quality problems. By analyzing the pixel features in the product feedback images uploaded by users and combining the image content (such as product usage status, environment, etc.), user satisfaction can be indirectly inferred. For example, frequently appearing dim or overexposed images may suggest that users have a poor experience with the product. Using pixel values ​​and image processing techniques (e.g., edge detection, texture analysis) to identify specific product features in an image can help categorize user feedback, differentiate between different models or versions of a product, or identify which part of a product user feedback is focused on.

[0053] The user evaluation data are classified, including user comments on each product at each time monitoring point, user ratings of each product, and page views of user comments on each product, and a user feedback quality evaluation index at each time monitoring point is obtained through comprehensive analysis.

[0054] Specifically, the user evaluation data is classified, and the specific process is as follows: extract the user score of each product from the user evaluation data, compare each product user score with the reference critical score stored in the product database, if a product user score is greater than or equal to the reference critical score, then mark the product user comment corresponding to the product user score as a positive product comment, and count the total number of product user comments and the number of positive product comments at each time monitoring point.

[0055] Furthermore, a comprehensive analysis is performed to obtain the user feedback quality evaluation index at each time monitoring point. The specific analysis process is as follows: the total page views of product user comments and the page views of product positive comments at each time monitoring point are collected, the total page views of product user comments and the page views of product positive comments at each time monitoring point are calculated by ratio, and the ratio result is marked as the positive comment influencing factor;

[0056] Based on the user ratings of each product at each time monitoring point, the total number of product user reviews, the number of positive product reviews and the positive review influencing factors, a comprehensive analysis is conducted to obtain the user feedback quality evaluation index at each time monitoring point.

[0057] In a specific embodiment, the numerical expression of the user feedback quality evaluation index at each time monitoring point is:

[0058]

[0059] where Gs represents the user feedback quality evaluation index at the s-th time monitoring point, Ps1 represents the number of positive product reviews at the s-th time monitoring point, Ps represents the total number of product user reviews at the s-th time monitoring point, ω represents the positive review influencing factor, i represents the ID of each product user, i = 1, 2, 3, ..., n, n represents the total number of product users, Fs→i represents the i-th user score at the s-th time monitoring point, and δ represents the feedback quality correction factor corresponding to the set user score.

[0060] Table 1 Example of user feedback quality evaluation index data

[0061]

[0062] It should be explained that the user ratings in Table 1 are the average scores of 100 users. For the convenience of calculation, it can be seen from Table 1 that when the user ratings increase, the user feedback quality evaluation index also increases. The larger the ratio of the number of positive product reviews to the total number of product user reviews, the larger the user feedback quality evaluation index. Both are positively correlated.

[0063] Specifically, the user feedback quality evaluation index at each time monitoring point is a quantitative evaluation data obtained by comprehensively analyzing the user ratings of each product, the total number of product user reviews, the number of product positive reviews and the positive review influencing factors at each time monitoring point, and is used to quantitatively evaluate the quality of user feedback on the target product.

[0064] Furthermore, in this embodiment, the value range of the positive comment impact factor is [0,1]. The positive comment impact factor reflects the ratio between the total number of views of product user comments and the number of views of product positive comments at each time monitoring point, and is used to reflect the importance of the number of positive comments relative to the total number of comments. Through this factor, we can emphasize the influence of positive comments in the overall feedback. The value range of the feedback quality correction factor corresponding to the user score is [0.1,0.5]. The feedback quality correction factor corresponding to the user score reflects the different contributions of different scores to the feedback quality. This means that different scores may have different proportions of impact on the final evaluation index. For example, a very satisfactory high score may be given more weight through a larger feedback quality correction factor value corresponding to the user score, while a lower score may reduce its impact on the total evaluation value through a smaller feedback quality correction factor value corresponding to the user score, which is used to reduce calculation errors and improve calculation accuracy.

[0065] It should be explained that, in addition to being calculated using the above formula, the user feedback quality evaluation index at each time monitoring point in this embodiment can also be obtained by using natural language processing technology to perform sentiment analysis and topic extraction on the user feedback text, and identify the proportion of positive and negative feedback. It can also be obtained by using tools to automatically collect and organize feedback data, training models to automatically classify feedback (such as complaints, suggestions, and praise) data, and predicting the priority and value of feedback.

[0066] It should be explained that, in this embodiment, the user ratings of each product at each time monitoring point, the total number of product user comments, and the number of positive product comments are obtained by processing all user comment data before the current time monitoring point. The user feedback quality evaluation index at each time monitoring point is used to quantify the user's satisfaction with the product at a specific time point or the overall quality of feedback. The larger the value, the more positive the user feedback and the better the product quality perception. The number of positive product comments directly reflects the number of positive feedback and is the basis for measuring the product's favorable reviews. More positive comments usually mean higher user satisfaction. This formula aims to quantify and evaluate the user feedback quality of the product at a specific time point by comprehensively considering the positive feedback ratio of the product, the individual user's rating and its relative importance through a mathematical model. It not only focuses on the absolute number of positive comments, but also deeply analyzes the distribution of user ratings, and adjusts the impact of different ratings by introducing correction factors, thereby obtaining a more comprehensive and detailed quality evaluation index. This has important reference value for product development, market strategy adjustment or service quality improvement.

[0067] A comprehensive analysis is conducted on the product visual quality evaluation index and user feedback quality evaluation index at each time monitoring point to obtain the product satisfaction evaluation index parameters at each time monitoring point. Based on the product satisfaction evaluation index parameters at each time monitoring point, user satisfaction is analyzed to obtain the analysis results and provide feedback.

[0068] Specifically, the product satisfaction evaluation index parameters at each time monitoring point, the specific analysis process is: comprehensive analysis of the product visual quality evaluation index and the user feedback quality evaluation index at each time monitoring point, to obtain the product satisfaction evaluation index parameters, the product satisfaction evaluation index parameters are used to quantitatively evaluate the user's satisfaction with the target product.

[0069] In a specific embodiment, the product satisfaction evaluation index parameter at each time monitoring point is specifically expressed as follows:

[0070]

[0071] In the formula, represents the product satisfaction evaluation index parameter at the sth time monitoring point, softplus(x)=ln(1+ex ), this function is usually used to construct activation functions to introduce nonlinear transformations and help the model learn more complex features and patterns. It is widely used in various network architectures, such as convolutional neural networks, recurrent neural networks, and generative adversarial networks. Ys represents the product visual quality evaluation index at the s-th time monitoring point, Gs represents the user feedback quality evaluation index at the s-th time monitoring point, ε1 represents the satisfaction evaluation influence factor corresponding to the set visual quality evaluation index, and ε2 represents the satisfaction evaluation influence factor corresponding to the set user feedback quality evaluation index.

[0072] Specifically, in this embodiment, the value range of the satisfaction assessment influence factor corresponding to the visual quality assessment index and the satisfaction assessment influence factor corresponding to the user feedback quality assessment index is [0.4, 0.6]. The satisfaction assessment influence factor corresponding to the visual quality assessment index and the satisfaction assessment influence factor corresponding to the user feedback quality assessment index reflect the relative importance of the visual quality assessment index and the user feedback quality assessment index in the satisfaction assessment. By adjusting the values ​​of the satisfaction assessment influence factor corresponding to the visual quality assessment index and the satisfaction assessment influence factor corresponding to the user feedback quality assessment index, personalized user satisfaction evaluation can be performed on the target products at different time monitoring points.

[0073] It needs to be explained that Figure 3 As shown, curve a represents the relationship between the visual quality evaluation index and the product satisfaction evaluation index parameter when the user feedback quality evaluation index is 0.5, curve b represents the relationship between the visual quality evaluation index and the product satisfaction evaluation index parameter when the user feedback quality evaluation index is 1.0, and curve c represents the relationship between the visual quality evaluation index and the product satisfaction evaluation index parameter when the user feedback quality evaluation index is 1.5. Figure 3 It can be obtained that the parameter of the product satisfaction evaluation index increases with the increase of the visual quality evaluation index. When the user feedback quality evaluation index becomes larger and larger, the parameter of the product satisfaction evaluation index also becomes larger and larger. The former is positively correlated with the latter.

[0074] It should be explained that, in addition to being calculated by the above formula, the product satisfaction evaluation index parameters at each time monitoring point in this embodiment can also be obtained by using social media monitoring tools (such as Hootsuite, Brandwatch) to monitor brand mentions and product evaluations, and analyzing user emotional tendencies and discussion hotspots. It can also be obtained by tracking user behaviors such as usage frequency, function click-through rate, dwell time, etc. through analysis tools (such as Google Analytics, Firebase), indirectly reflecting user satisfaction, monitoring user activity, retention rate and churn rate, and evaluating users' continued satisfaction and loyalty to the product.

[0075] It should be explained that in this embodiment, product visual quality and user feedback quality are two important dimensions for evaluating user satisfaction with the product, and they interact with each other from different aspects to affect the user's overall feelings and evaluation. User expectations of product feedback images may affect their ratings of products. If the image quality meets or exceeds user expectations, users may give higher ratings; conversely, if the image quality is lower than user expectations, users may give lower ratings even if the actual experience is good. Visual quality is the first step to attract users and determines whether users are willing to learn more about the product; and user feedback quality is the basis for continuous optimization and improvement, ensuring that the product can meet user needs in the long term and enhance user loyalty.

[0076] Specifically, the user satisfaction is analyzed, the analysis results are obtained and feedback is provided, and the specific analysis process is: curve fitting is performed on the product satisfaction evaluation index parameters at each time monitoring point through polynomial regression to obtain the time series curve and time-dependent function of the product satisfaction evaluation index parameters at each time monitoring point, the most recent time monitoring point is marked as the current time monitoring point, the time-dependent function of the product satisfaction evaluation index parameters at the current time monitoring point is derived, the time derivative of the product satisfaction evaluation index parameters at the current time monitoring point is obtained, and it is marked as a reference index for the satisfaction change trend;

[0077] Compare the satisfaction change trend reference index with 0. If the satisfaction change trend reference index is greater than or equal to 0, the product user's satisfaction change trend is marked as a stable trend. If the satisfaction change trend reference index is less than 0, the product user's satisfaction change trend is marked as a downward trend.

[0078] The analysis results are fed back to relevant personnel, who then improve the product based on the actual content of user comments.

[0079] Specifically, feedback can be provided to relevant personnel by writing an analysis report. First, at the beginning of the report, briefly introduce the purpose of the analysis, the methods used (polynomial regression, curve fitting, etc.), the data source (satisfaction evaluation indicators at each time monitoring point), and the main findings of the analysis, especially the reference index of the satisfaction change trend at the current time point and its meaning. Then use charts to intuitively display the time series curve, including a chart of the product satisfaction change trend over time, a fitting curve, and a clear mark of the derivative at the current time point (i.e., the reference index of the satisfaction change trend). Explain in detail the reference index of the satisfaction change trend at the current time point, explain what a stable trend greater than or equal to 0 means, what a downward trend less than 0 means, and the impact these trends may have on the product and user base to reduce the difficulty of understanding the report for relevant personnel.

[0080] In a specific embodiment, the product database refers to a target product database, which is usually created for a specific product or service to store, manage and analyze a collection of various data related to the product. This database may include but is not limited to product specifications, user feedback, sales data, market performance, competitor analysis, customer behavior data, etc.

[0081] Reference Figure 2 As shown, the second aspect of the present invention provides a satisfaction evaluation system based on network data capture, including: a network data acquisition module, an image feedback data processing module, a user evaluation data analysis module, a comprehensive analysis module and a product database.

[0082] Network data collection module, used to capture network evaluation data of target products;

[0083] An image feedback data processing module is used to process the image feedback data of the target product to obtain the product visual quality evaluation index at each time monitoring point;

[0084] The user evaluation data analysis module is used to obtain the user feedback quality evaluation index at each time monitoring point through comprehensive analysis based on user ratings, the number of user comments in each category, and the number of comment views;

[0085] Comprehensive analysis module, used to comprehensively analyze the product visual quality evaluation index and user feedback quality evaluation index at each time monitoring point, and obtain the satisfaction evaluation index parameters of the target product. According to the satisfaction evaluation index parameters of the target product, the target product is evaluated, the evaluation results are obtained and feedback is given;

[0086] The product database is used to store relevant evaluation data of the target product, including the reference standard image and the pixel values ​​of its corresponding pixel points, the allowable deviation pixel values, the reference critical score and the reference index of the satisfaction change trend.

[0087] The above contents are merely examples and explanations of the structure of the present invention. The technicians in this technical field may make various modifications or additions to the specific embodiments described or replace them in a similar manner. As long as they do not deviate from the structure of the invention or exceed the scope defined by the claims, they should all fall within the protection scope of the present invention.

Claims

1. A satisfaction evaluation method based on network data crawling, characterized in that: include: Capturing network evaluation data of a target product, wherein the network evaluation data of the target product includes image feedback data and user evaluation data of the target product; Process the image feedback data of the target product to obtain the product visual quality evaluation index at each time monitoring point; Classify the user evaluation data, including user comments on each product at each time monitoring point, user ratings of each product, and page views of user comments on each product, and comprehensively analyze to obtain a user feedback quality evaluation index at each time monitoring point; Comprehensively analyze the product visual quality evaluation index and user feedback quality evaluation index at each time monitoring point to obtain the product satisfaction evaluation index parameters at each time monitoring point. According to the product satisfaction evaluation index parameters at each time monitoring point, analyze the user satisfaction, obtain the analysis results and provide feedback; The specific numerical expression of the product visual quality evaluation index at each time monitoring point is: ； In the formula, represents the product visual quality evaluation index at the s-th time monitoring point, s represents the number of each time monitoring point, , p represents the total number of time monitoring points, m represents the number of each new product feedback image, , t represents the total number of new product feedback images, r represents the number of each pixel, , h represents the total number of pixels, represents the pixel value of the rth pixel of the image of the mth target product at the sth time monitoring point, Represents the pixel value of the rth pixel of the reference standard image, Indicates the allowable deviation pixel value, Indicates the set visual quality assessment correction factor; The user satisfaction is analyzed, the analysis results are obtained and feedback is given, and the specific analysis process is as follows: The product satisfaction evaluation index parameters at each time monitoring point are curve-fitted by polynomial regression to obtain the time series curve of the product satisfaction evaluation index parameters at each time monitoring point and its time-dependent function. The most recent time monitoring point is marked as the current time monitoring point, and the time-dependent function of the product satisfaction evaluation index parameters at the current time monitoring point is derived to obtain the time derivative of the product satisfaction evaluation index parameters at the current time monitoring point, which is marked as a reference index for the satisfaction change trend. Compare the satisfaction change trend reference index with 0. If the satisfaction change trend reference index is greater than or equal to 0, the product user's satisfaction change trend is marked as a stable trend. If the satisfaction change trend reference index is less than 0, the product user's satisfaction change trend is marked as a downward trend. The analysis results are fed back to relevant personnel, who then improve the product based on the actual content of user comments.

2. A satisfaction evaluation method based on network data crawling according to claim 1, characterized in that: The image feedback data of the target product is processed to obtain the product visual quality evaluation index at each time monitoring point. The specific processing process is as follows: Deploy several time monitoring points, collect feedback images of new products at each time monitoring point, overlap and compare the feedback images of each new product at each time monitoring point with the reference standard images in the product database, extract the pixel values ​​of each pixel point of the feedback image of each new product at each time monitoring point and the corresponding pixel points of the reference standard image, and comprehensively analyze to obtain the product visual quality evaluation index at each time monitoring point.

3. The method for evaluating satisfaction based on network data capture according to claim 1, characterized in that: The specific process of classifying the user evaluation data is as follows: The user rating of each product is extracted from the user evaluation data, and each product user rating is compared with the reference critical rating stored in the product database. If a product user rating is greater than or equal to the reference critical rating, the product user review corresponding to the product user rating is marked as a positive product review, and the total number of product user reviews and the number of positive product reviews at each time monitoring point are counted.

4. A satisfaction evaluation method based on network data capture according to claim 3, characterized in that: The comprehensive analysis obtains the user feedback quality evaluation index at each time monitoring point. The specific analysis process is as follows: Collect the total page views of product user reviews and the page views of product positive reviews at each time monitoring point, perform a ratio operation on the total page views of product user reviews and the page views of product positive reviews at each time monitoring point, and mark the ratio result as a positive review influencing factor; Based on the user ratings of each product at each time monitoring point, the total number of product user reviews, the number of positive product reviews and the positive review influencing factors, a comprehensive analysis is conducted to obtain the user feedback quality evaluation index at each time monitoring point.

5. A satisfaction evaluation method based on network data capture according to claim 4, characterized in that: The specific analysis process of the product satisfaction evaluation index parameters at each time monitoring point is as follows: A comprehensive analysis is performed on the product visual quality evaluation index and the user feedback quality evaluation index at each time monitoring point to obtain product satisfaction evaluation index parameters, which are used to quantitatively evaluate user satisfaction with the target product.

6. The method for evaluating satisfaction based on network data capture according to claim 2 is characterized in that: The product visual quality evaluation index at each time monitoring point is quantitative evaluation data obtained by comparing and analyzing the feedback images of each new product at each time monitoring point with the reference standard image, and is used to quantitatively evaluate the visual quality of the target product and provide a basis for evaluating the satisfaction of the target product.

7. The method for evaluating satisfaction based on network data capture according to claim 5 is characterized by: The specific numerical expression of the product satisfaction evaluation index parameter at each time monitoring point is: ； In the formula, represents the product satisfaction evaluation index parameter at the s-th time monitoring point, , represents the product visual quality assessment index at the s-th time monitoring point, represents the user feedback quality evaluation index at the s-th time monitoring point, It represents the satisfaction evaluation impact factor corresponding to the set visual quality evaluation index. Indicates the satisfaction evaluation impact factor corresponding to the set user feedback quality evaluation index.

8. A system using a satisfaction evaluation method based on network data capture as described in any one of claims 1 to 7, characterized in that: include: Network data collection module, used to capture network evaluation data of target products; An image feedback data processing module is used to process the image feedback data of the target product to obtain the product visual quality evaluation index at each time monitoring point; The user evaluation data analysis module is used to obtain the user feedback quality evaluation index at each time monitoring point through comprehensive analysis based on user ratings, the number of user comments in each category, and the number of comment views; Comprehensive analysis module, used to comprehensively analyze the product visual quality evaluation index and user feedback quality evaluation index at each time monitoring point, and obtain the satisfaction evaluation index parameters of the target product. According to the satisfaction evaluation index parameters of the target product, the target product is evaluated, the evaluation results are obtained and feedback is given; The product database is used to store the online evaluation data of the target product, including the reference standard image and the pixel values ​​of its corresponding pixel points, the allowable deviation pixel values, the reference critical score and the reference index of the satisfaction change trend.

Images