User Demand Mining Of C Company’s New Energy Vehicle Products Based On Online Comments

With the improvement of people’s income and the progress of technology,more and more people choose to buy new energy vehicles as a means of travel.When buying a car,people often collect information through various channels to help them make decisions about buying a car.With the rapid development of Internet technology,a large number of professional car forums have emerged,in which you can easily see a large number of user comments and feelings,which provides buyers with convenient and reliable information sources.More and more car buyers will choose to collect and learn about related car conditions through online forums.From the perspective of production enterprises,it is more necessary to maintain the perspective of development,pay attention to the needs of users and reflected problems on the automobile forum,collect and analyze in time,and guide the improvement of products,which is of great significance for the automobile production enterprises to improve the competitiveness of products.This paper chooses C brand new energy vehicles as the research object,explores users’ needs by analyzing online comments,and puts forward improvement suggestions.Specific research contents are as follows:(1)Online comment collection.Web crawler technology was used to collect review data of new energy vehicles from Pacific Auto,Autohome,Aika Auto,Car quality net and other websites,mainly including review information of typical new energy vehicle products of brand C and competing models of the research object.(2)Standardized processing of comment text.For the text data normalization processing,word segmentation to stop words,in order to be more accurate segmentation,this paper constructs the automobile dictionary.By TFIDF method statistics,preliminary analysis comments.(3)Emotion analysis of online comments.First,based on the statement level sentiment analysis,the method in Snow NLP package was used to carry out the statement level sentiment analysis for the word segmentation results.The comments of appearance,interior,space,configuration,power,handling,comfort,shortcomings and other dimensions of each comment were scored.Within the range of 0 to 1,the score was greater than or equal to 0,5 was positive evaluation,and less than 0.5 was negative evaluation.Secondly,based on aspect-level sentiment analysis,LTP,a natural language processing tool of HIT,is used to conduct aspect-level sentiment analysis on online comments to form quantitative user opinion pairs.(4)LDA topic analysis.LDA model is used to extract the negative comments from the dimensions of appearance,interior decoration,space,configuration,power,control,comfort and shortcomings,and the positive comments from the dimensions of appearance,interior decoration and advantages,and find out the key issues and corresponding opinion pairs in each dimension.(5)Classification and improvement suggestions based on KANO requirements.Based on the quantitative data of sentiment analysis,the paper constructs a KANO demand classification model,comprehensively considers security,user emotion,technical difficulty and other factors,and finds out the priority order of demand improvement and puts forward improvement strategies under the condition of limited funds,time and personnel.So that the product close to the user needs,so that the investment is reported.This paper analyzes online reviews of new energy vehicles through text mining technology,analyzes and mines consumers’ emotional evaluation of new energy vehicles,further extracts consumers’ focus and product shortcomings,and provides improvement solutions based on user needs for new energy vehicle manufacturers through modeling analysis.Finally,the paper points out the possible shortcomings and problems that need further consideration.