User Demand Analysis Based On Text Mining

The enterprise's product planning and product positioning first need to obtain the real needs of users,accurate acquisition of the needs of users can occupy a favorable position in the market,is the decisive factor for the success of the product.With the rapid development of Internet technology,a huge amount of user-generated data will be generated every day on the online shopping platform,which contains a large number of user preferences,user purchase behavior and user demand information for products.This information has gradually become an extremely important data resource for user demand research.There are still many shortcomings in the traditional research on user demand acquisition.This paper takes the comment data of huawei Mate30 series mobile phone users as the research content,and combines the text data mining technology with Kano model to analyze the real needs of users,which is of great significance to the product improvement and improvement of enterprises.First of all,this paper expounds the background and significance of studying users' online comments in the Internet era,and summarizes the research purpose of this paper according to the background and significance of the research.The paper also introduces the relevant literature review from four aspects: the usefulness of users' online comments,feature extraction,emotion analysis and the study of users' needs of online comments.Secondly,relevant theories applied in this paper are studied,including web crawler technology,text mining technology,VSM classification algorithm,k-means clustering algorithm,Kano model and the general flow of user demand Kano model,to provide a theoretical basis for subsequent research.Then,the data of users' online comments on mobile phones of huawei Mate30 series in taobao.com and jingdong mall were taken as the experimental data source,using the web crawler technology for online reviews text data,first for data preprocessing,use in Python jieba segmentation tools,combined with the stop libraries to segmentation processing of data,the TF-IDF algorithm to build product key library,it is concluded that the user product characteristics to the attention of a direction mainly in the system,network,calls,text messages,battery,appearance,screen,camera,memory,entertainment,system,brand,buttons,service,etc.,The current comprehensive zhiwang emotion dictionary is used as the construction of themain emotion lexicon in this paper,and combined with the Chinese emotion lexicon of dalian university of technology,the simplified Chinese emotion polarity dictionary NTUSD as a supplement,the construction of the emotion lexicon,and the feature lexicon match for emotion analysis,analysis of the user's emotional expression of the product.Finally,by building the Kano model user demand,determine the level of user online review analysis indicators for the depth of the contents of the user reviews and comments on the influence,and for each primary index of each secondary index K-Means cluster analysis,combined with the Kano model elements: basic needs,expectations,charismatic,needs,and the connotation of independent demand,analyze the user's actual demand,the experimental results show that the user needs the Kano model to the user's basic needs,expectations,charismatic demand and independent demand can effectively identify.And to carry on the analysis to the above demand,put forward the product improvement suggestion.