Research On User Demand Identification And Classification Based On Online Review Minin

With the rapid development of the Internet,social networks and e-commerce have also developed rapidly.More and more users choose to make purchases online and share their product or service experience in the form of online comments after the shopping decision.Online comments have also become an important information carrier to discover users’ needs.It has also become a trend to mine the data information of e-commerce platform.How to effectively extract valuable information from the existing massive data information,obtain the most real needs of users,and promote enterprises to make more efficient and scientific management decisions is an important research issue.Therefore,this thesis selects the online comment information of users of four popular real lipstick products from the e-commerce platform of JD.com as the data source to analyze user needs.This thesis combines comment text analysis and Kano model to classify user needs reasonably,which can make up for the shortcomings of traditional user needs research.In the textual analysis of comments.The author uses LDA topic mining and sentiment analysis based on sentiment dictionary to analyze online comment information.First,LDA is used to obtain keywords from online reviews,and product features contained in the review text are found by filtering the keywords to find the main needs of users.According to the product characteristic attributes extracted from the text analysis and the inherent physical characteristic attributes given by the product,the subject keywords are summarized and sorted out.To realize the identification of seven main types of user needs.Secondly,emotion dictionary is used to analyze the comment information.Dependency syntax technology is used to extract feature-emotion word pairs,construct feature word database and emotion word database,identify the subjective sentences containing product features and emotion words,and analyze the emotional tendency of each subjective sentence to obtain the emotional tendency corresponding to different needs.User requirement classification based on Kano model.According to the classification study of Kano model,most scholars still adopt the form of questionnaire survey.The author tries to start from the original concept of Kano model and actual comment data,and classifies and analyzes different needs of users in terms of user satisfaction and user attention.The classification model of Kano model is constructed based on the proportion and importance of product features as well as the results of subjective sentence emotional tendency,and user needs are classified into necessary needs,expected needs,charm needs and undifferentiated needs.The study found that "product efficacy" and "product quality" are essential requirements."Product price" and "product color" belong to the charm demand."Product Appearance ","Additional Products" and "Services" are undifferentiated requirements.Based on the analysis results of Kano model,it puts forward targeted improvement strategies,and provides decision-making suggestions for the enterprise’s product update and iteration.