E-commerce platform user feedback analysis method and device, and medium

Abstract

The invention relates to an e-commerce platform user feedback analysis method and device and a medium, and relates to the technical field of e-commerce platforms, and the method comprises the steps: obtaining a plurality of pieces of user feedback information; converting the user feedback information into text data, and constructing a feedback text data set; extracting keywords according to the feedback text data set; counting the occurrence frequency of each keyword in the feedback text data set, and obtaining a keyword analysis result according to the occurrence frequency; calculating the information entropy of the text data of each piece of user feedback information, and obtaining a priority analysis result according to the information entropy; determining topic features and emotion features of the user feedback information according to the priority analysis result and the keyword analysis result; and obtaining a user feedback analysis result according to the topic features and the emotion features. The method has the effect of improving the accuracy and efficiency of user feedback analysis.

Description

Technical Field

[0001] The present application relates to the technical field of e-commerce platforms, and in particular to a method, device and medium for analyzing user feedback on e-commerce platforms. Background Art

[0002] An e-commerce platform is a platform that provides online transactions for businesses or individuals. It uses the Internet as a network infrastructure to provide shared resources such as payment platforms, security platforms, and management platforms, so that businesses or individuals can conduct commercial activities at low cost and high efficiency. With the surge in the number of users, e-commerce platforms are faced with the challenge of how to effectively manage and utilize massive user feedback information. This feedback information is of great significance for improving product and service quality and enhancing user experience.

[0003] At present, the industry generally adopts a variety of methods to process user feedback information. For example, some e-commerce platforms use manual review teams to read and organize user comments one by one. Although this can ensure a high degree of accuracy, it is time-consuming, labor-intensive and costly. Some e-commerce platforms have introduced simple keyword search tools to automatically extract keywords from user feedback, but this method often fails to fully capture the user's intentions and emotions, resulting in inaccurate analysis results. In addition, some e-commerce platforms have tried to use machine learning algorithms to perform preliminary classification of user feedback, but there are still problems such as high misjudgment rate and incomplete classification.

[0004] The above-mentioned user feedback analysis methods all have certain defects in practical applications. The manual review method is inefficient and costly, and it is difficult to adapt to the needs of large-scale data processing. Although simple keyword search tools have improved the processing speed, they lack in-depth analysis capabilities and are prone to missing key information. Although the existing machine learning algorithms have a certain degree of intelligence, their accuracy and precision still need to be improved. Summary of the invention The purpose of this application is to provide an e-commerce platform user feedback analysis method, device and medium to effectively improve the accuracy and efficiency of user feedback analysis.

[0005] In the first aspect, the present application provides an e-commerce platform user feedback analysis method using the following technical solution: A method for analyzing user feedback on an e-commerce platform, comprising: Get some user feedback information; Convert user feedback information into text data and construct a feedback text dataset; Extract keywords based on the feedback text dataset; Count the frequency of occurrence of each keyword in the feedback text dataset and obtain the keyword analysis results based on the frequency of occurrence; Calculate the information entropy of the text data of each user feedback information, and obtain the priority analysis result based on the information entropy; Determine the topic features and sentiment features of user feedback information based on the priority analysis results and keyword analysis results; The user feedback analysis results are obtained based on topic features and sentiment features.

[0006] By adopting the above technical solutions, this application screens feedback information through keyword extraction and priority extraction, and then analyzes feedback information through topic features and sentiment features, and comprehensively uses natural language processing technology and machine learning algorithms to extract and analyze key features of user feedback information from multiple dimensions, thereby helping e-commerce platforms to better understand and respond to user needs, improve user experience and platform service quality. Compared with traditional manual review and rule-based automatic classification systems, this application has a higher degree of automation and more powerful analysis capabilities, and can process a large amount of user feedback information in a short time, greatly improving the accuracy and efficiency of user feedback analysis.

[0007] Furthermore, the obtaining of several pieces of user feedback information includes: Obtain explicit user feedback by collecting online surveys, collecting user reviews and ratings, and monitoring social media posts; Obtain implicit user feedback information by obtaining user purchase behavior, browsing history, and search history; The explicit user feedback information and the implicit user feedback information are cleaned and preprocessed respectively to obtain user feedback information.

[0008] By adopting the above technical solution, the present application ensures the extensiveness and accuracy of collecting user feedback information by synchronously acquiring explicit user feedback information and implicit user feedback information, thereby improving the extensiveness and accuracy of data for subsequent user feedback analysis.

[0009] Furthermore, extracting keywords based on the feedback text dataset specifically includes: Segment the text data in the feedback text data set into words according to natural language processing technology, and restore the obtained words to their basic forms; Remove the basic words in the vocabulary to obtain a vocabulary data set; Calculate the TF-IDF value of each word in the vocabulary dataset in the feedback text dataset; Sort the words in the vocabulary dataset according to the TF-IDF value from large to small; Select the words ranked first as keywords.

[0010] By adopting the above technical solution, the TF-IDF (Term Frequency-Inverse Document Frequency) used in this application is a statistical method used to evaluate the importance of words in the feedback writing data set, which can effectively improve the accuracy of keyword analysis.

[0011] Further, the calculation of the TF-IDF value of each word in the vocabulary data set in the feedback text data set specifically includes: The word frequency TF of a word is obtained by dividing the number of times a word appears in the text data by the total number of words in the text data. The inverse document frequency (IDF) is obtained by dividing the logarithm of the total number of words in the text data by the number of text data containing the word plus 1. Multiply the term frequency TF and inverse document frequency IDF of each word respectively to get the TF-IDF value of each word in the feedback text dataset.

[0012] By adopting the above technical solution, TF in this application represents word frequency, that is, the number of times a word appears in the writing data set, and IDF represents inverse document frequency, which reflects the proportion of text data containing the word in the total amount of writing data. This application uses the TF-IDF algorithm to effectively screen out high-frequency and important keywords, thereby improving the accuracy of keyword analysis.

[0013] Furthermore, the counting of the occurrence frequency of each keyword in the feedback text data set and obtaining the keyword analysis result according to the occurrence frequency specifically includes: Count the number of occurrences of each keyword in the feedback text dataset.

[0014] Divide the number of occurrences of each keyword by the total number of words in the feedback text dataset to get the frequency of occurrence of each keyword; Sort the keywords by their frequency of occurrence from large to small to get a word frequency table; According to the preset conditions, the keywords with a frequency higher than the threshold are filtered as the keyword analysis results.

[0015] By adopting the above technical solution, the present application obtains the keyword analysis result through the frequency of occurrence of keywords, determines the threshold according to preset conditions, and then filters out keywords with a frequency higher than the threshold, thereby determining the keywords that have the greatest impact on the feedback information and ensuring the accuracy of the keyword analysis.

[0016] Furthermore, determining the topic features of the user feedback information according to the priority analysis results and the keyword analysis results specifically includes: Comprehensively select important feedback information based on priority analysis results and keyword analysis results; Use clustering algorithms to classify similar feedback information and identify explicit topic features; Use topic models to identify implicit topic features in text data of important feedback information; The topic features of feedback information are obtained by combining explicit topic features and implicit topic features.

[0017] By adopting the above technical solution, the present application uses clustering algorithms and topic models to extract topic features of feedback information, and obtains explicit topic features and implicit topic features, which double guarantees to ensure the comprehensiveness and accuracy of feedback information extraction.

[0018] Furthermore, determining the emotional features of the user feedback information according to the priority analysis results and the keyword analysis results specifically includes: Identify the sentiment tendency in the feedback information text data according to the NRC vocabulary sentiment association dictionary; Emotional fluctuations are obtained based on the changing trend of emotional tendencies over time; The emotional characteristics of feedback information are obtained based on emotional tendencies and emotional fluctuations.

[0019] By adopting the above technical solution, the present application utilizes the NRC vocabulary sentiment association dictionary to extract the sentiment characteristics in the feedback information based on the sentiment characteristics, analyzes the user feedback according to the sentiment characteristics, and improves the accuracy and comprehensiveness of the analysis results.

[0020] Furthermore, before extracting keywords according to the feedback text dataset, the method further includes: Obtain the user identity information corresponding to each piece of user feedback information; Filter whether there is duplicate user feedback information based on user identity information; The duplicate user feedback information is removed to obtain a new feedback text dataset.

[0021] By adopting the above technical solution, the present application can improve the accuracy of the analysis results and the efficiency of the analysis process by screening repeated feedback information of the same user and reducing the amount of data for subsequent analysis.

[0022] In a second aspect, the present application provides a computer device that adopts the following technical solution: A computer device comprises a memory and a processor, wherein the memory stores a method for analyzing user feedback on an e-commerce platform as described in the first aspect, which can be loaded and executed by the processor.

[0023] In a second aspect, a computer-readable storage medium provided by the present application adopts the following technical solution: A computer-readable storage medium stores a computer program that can be loaded by a processor and executed by a method for analyzing user feedback on an e-commerce platform as described in the first aspect.

[0024] In summary, the present application includes at least one of the following beneficial technical effects: 1. This application screens feedback information through keyword extraction and priority extraction, and then analyzes the feedback information through topic features and sentiment features. It comprehensively uses natural language processing technology and machine learning algorithms to extract and analyze key features of user feedback information from multiple dimensions, thereby helping e-commerce platforms better understand and respond to user needs and improve user experience and platform service quality; 2. This application has a higher degree of automation and more powerful analysis capabilities, and can process a large amount of user feedback information in a short period of time, greatly improving the accuracy and efficiency of user feedback analysis; 3. This application ensures the extensiveness and accuracy of user feedback information collected by synchronously acquiring explicit user feedback information and implicit user feedback information, thereby improving the extensiveness and accuracy of data for subsequent user feedback analysis; 4. This application uses clustering algorithms and topic models to extract topic features of feedback information to obtain explicit topic features and implicit topic features, which provide dual guarantees to ensure the comprehensiveness and accuracy of feedback information extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a flow chart of a user feedback analysis method according to an embodiment of the present application; Figure 2 is a specific method flow chart of step S3 of the embodiment of the present application; Figure 3 This is a specific method flow chart of step S4 of an embodiment of the present application. DETAILED DESCRIPTION

[0026] The following is combined with Figure 1 -Attached Figure 3 , further details of this application are given.

[0027] At present, the industry generally adopts a variety of means to process user feedback information. For example, some e-commerce platforms use manual review teams to read and organize user comments one by one. Although this can ensure a high degree of accuracy, it is time-consuming, labor-intensive and costly. Some e-commerce platforms have introduced simple keyword search tools to automatically extract keywords from user feedback, but this method often fails to fully capture the user's intentions and emotions, resulting in inaccurate analysis results. In addition, some e-commerce platforms have tried to use machine learning algorithms to preliminarily classify user feedback, but there are still problems such as high misjudgment rate and incomplete classification. The above user feedback analysis methods all have certain defects in practical applications. The manual review method is inefficient and costly, and it is difficult to adapt to the needs of large-scale data processing; although the simple keyword search tool improves the processing speed, it lacks in-depth analysis capabilities and is prone to missing key information; and although the existing machine learning algorithms have a certain degree of intelligence, they still need to be improved in terms of accuracy and precision.

[0028] Therefore, the present application embodiment provides an e-commerce platform user feedback analysis method, referring to Figure 1 ,include: S1. Obtain several pieces of user feedback information.

[0029] Specifically, explicit user feedback information is obtained by collecting online questionnaires, collecting user comments and ratings, and monitoring social media speeches. Implicit user feedback information is obtained by obtaining user purchase behavior, browsing history, and search history. Data cleaning and preprocessing are performed on explicit user feedback information and implicit user feedback information to obtain user feedback information. In the specific implementation process, explicit user feedback information can be obtained by crawling user comments and ratings from e-commerce platforms or social media through crawler technology, or by connecting to the e-commerce platform server through an API interface to obtain user explicit feedback information in real time. Implicit user feedback information can be obtained by querying the user's historical purchase record and browsing record through the database.

[0030] S2. Convert user feedback information into text data and construct a feedback text dataset.

[0031] Specifically, natural language processing technology can be used to convert user feedback information into a standard text format, such as a UTF-8 encoded text file. In addition, the text data can also be segmented to split long sentences into short sentences for subsequent processing.

[0032] S3. Extract keywords based on the feedback text dataset.

[0033] Specifically, keywords that can reflect the characteristics of the text dataset are obtained according to the importance of the vocabulary in the text dataset for keyword analysis. In the specific implementation process, keywords can be determined by calculating the TF-IDF value. TF-IDF (Term Frequency-Inverse Document Frequency) is a statistical method used to evaluate the importance of vocabulary in the feedback writing dataset.

[0034] S4. Count the occurrence frequency of each keyword in the feedback text data set, and obtain the keyword analysis result based on the occurrence frequency.

[0035] Specifically, the keyword analysis result is obtained through the frequency of occurrence of the keyword, a threshold is determined according to a preset condition, and then the keywords with a frequency higher than the threshold are screened, thereby determining the keywords with the greatest impact on the feedback information.

[0036] S5. Calculate the information entropy of the text data of each user feedback information, and obtain a priority analysis result based on the information entropy.

[0037] Specifically, the Shannon entropy formula can be used to calculate the information entropy of the text data of each user feedback information, and the user feedback information can be sorted according to the size of the information entropy to obtain the priority analysis result. In the specific implementation process, the user feedback information with higher information entropy indicates that the content is more complex and needs to be processed first.

[0038] S6. Determine the topic features and sentiment features of the user feedback information based on the priority analysis results and the keyword analysis results.

[0039] Specifically, the user feedback information is comprehensively analyzed by extracting the topic features and sentiment features of the feedback information.

[0040] S7. Obtain user feedback analysis results based on topic features and sentiment features.

[0041] Specifically, a detailed user feedback report can be generated, including keyword analysis results, topic features and sentiment features, for reference by e-commerce platform operators. For example, the report can point out common user complaints and improvement suggestions for a product, helping operators to adjust product strategies in a timely manner.

[0042] The implementation principle of the embodiments of this application is as follows: By extracting keywords and priorities to screen the feedback information, and then analyzing the feedback information through topic features and sentiment features, comprehensively applying natural language processing technology and machine learning algorithms to extract and analyze the key features of user feedback information from multiple dimensions, so as to help the e-commerce platform better understand and respond to user needs, and improve user experience and platform service quality. Compared with traditional manual review and rule-based automatic classification systems, the embodiments of this application have a higher degree of automation and more powerful analysis capabilities, can process a large amount of user feedback information in a short time, and greatly improve the accuracy and efficiency of user feedback analysis.

[0043] As Figure 2 shown, step S3 of the embodiments of this application specifically includes: S31. According to natural language processing technology, split the text data in the feedback text dataset into words, and restore the obtained words to their basic forms.

[0044] Specifically, first convert the feedback text dataset into the same format, use natural language processing technology to split the text into a sequence of words or phrases, and restore the words to their basic forms to reduce lexical variants and improve analysis efficiency.

[0045] S32. Remove the basic words in the words to obtain a word dataset.

[0046] Specifically, remove the basic words in the text, such as "de", "he", "shi", etc. These words usually contribute little to the analysis results. In the specific implementation process, the NLTK library can be used for word segmentation and stop word removal.

[0047] S33. Calculate the TF-IDF value of each word in the word dataset in the feedback text dataset.

[0048] Specifically, TF-IDF (Term Frequency-Inverse Document Frequency) is a statistical method used to evaluate the importance of words in the feedback text dataset. In the specific implementation process, the Scikit-learn library can be used to calculate the TF-IDF value.

[0049] S34. Sort the words in the word dataset from largest to smallest according to the TF-IDF value.

[0050] S35. Select the top several words in the sorting as keywords.

[0051] Specifically, through the TF-IDF algorithm, high-frequency and important keywords can be effectively screened out, and the top several words in the sorting need to be selected as keywords according to the required number and accuracy of keywords.

[0052] Step S33 of the embodiment of the present application specifically includes: S331. Obtain the word frequency TF of the word by dividing the number of times the word appears in the text data by the total number of words in the text data.

[0053] Specifically, TF stands for word frequency, that is, the number of times a word appears in a text dataset. S332: Obtain the inverse document frequency (IDF) by dividing the logarithm of the total number of words in the text data by the number of text data containing the word plus 1.

[0054] Specifically, IDF stands for Inverse Document Frequency, which reflects the proportion of text data containing the word to the total amount of text data.

[0055] S333. Multiply the term frequency TF and the inverse document frequency IDF of each word respectively to obtain the TF-IDF value of each word in the feedback text dataset.

[0056] The implementation principle of the embodiment of the present application is: by using the TF-IDF algorithm, high-frequency and important keywords can be effectively screened out, thereby improving the accuracy of keyword analysis.

[0057] like Figure 3 As shown, step S4 of the embodiment of the present application specifically includes: S41. Count the number of occurrences of each keyword in the feedback text dataset.

[0058] S42. Divide the number of occurrences of each keyword by the total number of words in the feedback text data set to obtain the occurrence frequency of each keyword.

[0059] S43. Sort the keywords according to their frequency of occurrence from large to small to obtain a word frequency table.

[0060] S44. Filter keywords with a frequency higher than a threshold value according to preset conditions as keyword analysis results.

[0061] In the specific implementation process, keywords with an appearance frequency exceeding 0.01 may be set as important keywords to obtain keyword analysis results.

[0062] The implementation principle of the embodiment of the present application is: obtain the keyword analysis result through the frequency of occurrence of keywords, determine the threshold according to preset conditions, and then filter out keywords with a frequency higher than the threshold, so as to determine the keywords that have the greatest impact on the feedback information and ensure the accuracy of the keyword analysis.

[0063] In step S6 of the embodiment of the present application, determining the topic features of the user feedback information according to the priority analysis results and the keyword analysis results specifically includes: S61. Comprehensively select important feedback information based on the priority analysis results and keyword analysis results.

[0064] S62. Use a clustering algorithm to classify similar feedback information and identify explicit topic features.

[0065] S63. Use the topic model to identify implicit topic features in the text data of important feedback information; S64. Combining the explicit topic features with the implicit topic features to obtain the topic features of the feedback information.

[0066] In the specific implementation process, the K-means clustering algorithm can be used for clustering, and the LDA topic model can be used to identify implicit topic features.

[0067] The implementation principle of the embodiment of the present application is: using clustering algorithm and topic model to extract topic features of feedback information, to obtain explicit topic features and implicit topic features, double guarantee to ensure the comprehensiveness and accuracy of feedback information extraction.

[0068] In step S6 of the embodiment of the present application, determining the emotional features of the user feedback information according to the priority analysis results and the keyword analysis results specifically includes: S65. Identify the sentiment tendency in the feedback information text data according to the NRC vocabulary sentiment association dictionary.

[0069] S66. Emotional fluctuations are obtained based on the changing trend of emotional tendencies over time.

[0070] S67. Get emotional characteristics of feedback information based on emotional tendencies and emotional fluctuations.

[0071] In the specific implementation process, the TextBlob library can be used to perform sentiment analysis to identify positive, negative, and neutral sentiments.

[0072] The implementation principle of the embodiment of the present application is: using the NRC vocabulary sentiment association dictionary to extract the sentiment characteristics in the feedback information based on the sentiment characteristics, analyzing the user feedback based on the sentiment characteristics, and improving the accuracy and comprehensiveness of the analysis results.

[0073] In order to further enhance the accuracy of user feedback analysis, before extracting keywords from the feedback text dataset, the embodiment of the present application further includes: A1. Obtain the user identity information corresponding to each piece of user feedback information.

[0074] Specifically, the user's identity, such as user ID, user name, etc., can be obtained through the user login information of the e-commerce platform. This step helps to identify multiple feedbacks from the same user and avoid repeated processing.

[0075] A2. Filter whether there is duplicate user feedback information based on user identity information.

[0076] Specifically, the user identity information and the feedback content can be compared. If the same user submits the same feedback information in the same time period, it is considered as duplicate feedback. In the specific implementation process, the user feedback information can be fingerprinted using a hash function to compare whether the fingerprints are consistent.

[0077] A3. Remove duplicate user feedback information to obtain a new feedback text dataset.

[0078] Specifically, data deduplication technology, such as the DISTINCT keyword of SQL or the set data structure of Python, can be used to remove duplicate user feedback information. In the specific implementation process, the drop_duplicates method in the Pandas library can be used to remove duplicate rows. The implementation principle of this embodiment is: by increasing the acquisition and deduplication steps of user identity information, user feedback information can be processed and analyzed more effectively, avoiding resource waste and analysis errors caused by repeated feedback. This is of great significance for improving the accuracy and reliability of the analysis results, especially when processing large-scale user feedback data, it can significantly improve the robustness and efficiency of the system.

[0079] In order to eliminate the influence of text length differences on information entropy, when calculating the information entropy of the text data of each user feedback information, the embodiment of the present application also includes: normalizing the length of the text data. Specifically, the length of the text data of each user feedback information can be normalized so that texts of different lengths are comparable in the information entropy calculation. For example, the maximum and minimum normalization method can be used to normalize the text length to the interval [0, 1]. The implementation principle of this embodiment is: by normalizing the text length, the information entropy of user feedback information of different lengths can be evaluated more fairly, avoiding analysis bias caused by differences in text length. This plays an important role in improving the fairness and accuracy of the analysis results, especially when processing user feedback information of different lengths, it can ensure that the calculation results of the information entropy are more representative.

[0080] In order to identify the time trend and periodic characteristics of user feedback, when determining the topic characteristics of user feedback information according to the priority analysis results and keyword analysis results, the embodiment of the present application also includes: analyzing the time dimension of the feedback information. Specifically, the timestamp of the user feedback information can be included in the analysis scope to identify the time trend and periodic characteristics of the user feedback. In the specific implementation process, time series analysis methods, such as ARIMA models, can be used to identify seasonal changes and long-term trends in user feedback. The implementation principle of this embodiment is: by adding the analysis of the time dimension, it is possible to more comprehensively understand the time distribution law of user feedback, identify the seasonal and periodic characteristics of user feedback, and thus help the e-commerce platform better predict and respond to changes in user demand. This is of great significance for improving the dynamic response capability and market competitiveness of the platform.

[0081] The embodiment of the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program that can be loaded by the processor and executed by a method for analyzing user feedback on an e-commerce platform, specifically including: S1. Obtain several pieces of user feedback information.

[0082] S2. Convert user feedback information into text data and construct a feedback text dataset.

[0083] S3. Extract keywords based on the feedback text dataset.

[0084] S4. Count the frequency of occurrence of each keyword in the feedback text dataset, and obtain the keyword analysis results based on the frequency of occurrence. S5. Calculate the information entropy of the text data of each user feedback information, and obtain the priority analysis result according to the information entropy. S6. Determine the topic features and sentiment features of the user feedback information based on the priority analysis results and the keyword analysis results.

[0085] S7. Obtain user feedback analysis results based on topic features and sentiment features.

[0086] Wherein, the computer is a server, and the computer device includes a processor, a memory, a network interface and a database connected through a system bus. Wherein, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store user feedback information, feedback text data sets, keyword analysis results, priority analysis results, etc. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a user feedback analysis method for an e-commerce platform can be implemented.

[0087] The present application also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: S1. Obtain several pieces of user feedback information.

[0088] S2. Convert user feedback information into text data and construct a feedback text dataset.

[0089] S3. Extract keywords based on the feedback text dataset.

[0090] S4. Count the frequency of occurrence of each keyword in the feedback text dataset, and obtain the keyword analysis results based on the frequency of occurrence. S5. Calculate the information entropy of the text data of each user feedback information, and obtain the priority analysis result according to the information entropy. S6. Determine the topic features and sentiment features of the user feedback information based on the priority analysis results and the keyword analysis results.

[0091] S7. Obtain user feedback analysis results based on topic features and sentiment features.

[0092] In addition, the processor in the computer device executes all the steps of the above-mentioned e-commerce platform user feedback analysis method when executing the computer program.

[0093] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM) and memory bus dynamic RAM (RDRAM).

[0094] Those skilled in the art can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0095] The embodiments of this specific implementation are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. The same components are represented by the same figure marks. Therefore, any equivalent changes made based on the structure, shape, and principle of the present application should be included in the protection scope of the present application.

Claims

1. A method for analyzing user feedback on an e-commerce platform, characterized in that: include: Get some user feedback information; Convert user feedback information into text data and construct a feedback text dataset; Extract keywords based on the feedback text dataset; Count the frequency of occurrence of each keyword in the feedback text dataset, and obtain the keyword analysis results based on the frequency of occurrence; Calculate the information entropy of the text data of each user feedback information, and obtain the priority analysis result based on the information entropy; Determine the topic features and sentiment features of user feedback information based on the priority analysis results and keyword analysis results; The user feedback analysis results are obtained based on topic features and sentiment features.

2. The method for analyzing user feedback on an e-commerce platform according to claim 1, characterized in that: The obtaining of several pieces of user feedback information includes: Obtain explicit user feedback by collecting online surveys, collecting user reviews and ratings, and monitoring social media posts; Obtain implicit user feedback information by obtaining user purchase behavior, browsing history, and search history; The explicit user feedback information and the implicit user feedback information are cleaned and preprocessed respectively to obtain user feedback information.

3. The method for analyzing user feedback on an e-commerce platform according to claim 1, characterized in that: The extracting keywords according to the feedback text dataset specifically includes: Segment the text data in the feedback text data set into words according to natural language processing technology, and restore the obtained words to their basic forms; Remove the basic words in the vocabulary to obtain a vocabulary data set; Calculate the TF-IDF value of each word in the vocabulary dataset in the feedback text dataset; Sort the words in the vocabulary dataset according to the TF-IDF value from large to small; Select the words ranked first as keywords.

4. The method for analyzing user feedback on an e-commerce platform according to claim 3, characterized in that: The calculating of the TF-IDF value of each word in the vocabulary data set in the feedback text data set specifically includes: The word frequency TF of a word is obtained by dividing the number of times a word appears in the text data by the total number of words in the text data. The inverse document frequency (IDF) is obtained by dividing the logarithm of the total number of words in the text data by the number of text data containing the word plus 1. Multiply the term frequency TF and inverse document frequency IDF of each word respectively to get the TF-IDF value of each word in the feedback text dataset.

5. The method for analyzing user feedback on an e-commerce platform according to claim 1, characterized in that: The counting of the occurrence frequency of each keyword in the feedback text data set and obtaining the keyword analysis result according to the occurrence frequency specifically includes: Count the number of times each keyword appears in the feedback text dataset; Divide the number of occurrences of each keyword by the total number of words in the feedback text dataset to get the frequency of occurrence of each keyword; Sort the keywords by their frequency of occurrence from large to small to get a word frequency table; According to the preset conditions, the keywords with a frequency higher than the threshold are filtered as the keyword analysis results.

6. The method for analyzing user feedback on an e-commerce platform according to claim 1, characterized in that: Determining the topic features of the user feedback information according to the priority analysis results and the keyword analysis results specifically includes: Comprehensively select important feedback information based on priority analysis results and keyword analysis results; Use clustering algorithms to classify similar feedback information and identify explicit topic features; Use topic models to identify implicit topic features in text data of important feedback information; The topic features of feedback information are obtained by combining explicit topic features and implicit topic features.

7. The method for analyzing user feedback on an e-commerce platform according to claim 6, characterized in that: Determining the emotional features of the user feedback information according to the priority analysis results and the keyword analysis results specifically includes: Identify the sentiment tendency in the feedback information text data according to the NRC vocabulary sentiment association dictionary; Emotional fluctuations are obtained based on the changing trend of emotional tendencies over time; The emotional characteristics of feedback information are obtained based on emotional tendencies and emotional fluctuations.

8. The method for analyzing user feedback on an e-commerce platform according to any one of claims 1 to 7, characterized in that: Before extracting keywords from the feedback text dataset, it also includes: Obtain the user identity information corresponding to each piece of user feedback information; Filter whether there is duplicate user feedback information based on user identity information; The duplicate user feedback information is removed to obtain a new feedback text dataset.

9. A computer device, characterized in that: It includes a memory and a processor, wherein the memory stores a method for analyzing user feedback on an e-commerce platform that can be loaded and executed by the processor as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: A computer program is stored which can be loaded by a processor and executes an e-commerce platform user feedback analysis method as described in any one of claims 1 to 8.

Images