Text Mining Analysis Of Online User Review

In recent years,with the rapid development and popularization of e-commerce,consumers' consumption habits have undergone profound and tremendous changes.A considerable part of consumption and trading behavior has shifted from offline to online,and the scale of online transactions has been expanding.In 2017,China's online retail market reached 720 million yuan,and as of December 2018,China's online shopping users reached 610 million yuan.Consumers can publish their own evaluation of purchased goods through e-commerce platform,which generates a lot of evaluation information.This not only provides reference for other consumers'purchase,but also provides a lot of valuable information for businesses.Potential consumers can get a comprehensive and detailed understanding of the goods or services they are concerned about through online user evaluation information,and make further decisions on whether to buy them.Businessmen can learn consumers'consumption habits,characteristics of interest and consumption intentions from the evaluation information,so as to improve products and services and enhance the competitiveness of enterprises.Therefore,it is of great significance to study and analyze user's comment data.In this paper,two text mining methods,LDA model and emotional tendency analysis,are used to conduct an in-depth analysis of the comment text,taking the mi 8 phone in the flagship store of xiaomi jingdong as an example.The specific content is as follows:first,the web crawler technology is used to achieve the acquisition of mobile phone comment text,and data preprocessing is realized through word segmentation,word stopping and other steps.Secondly,the LDA model is used to classify all the comment texts into 20 themes.It is learned that users focus on the eight aspects of the mi 8 phone,including shopping experience,phone performance,phone price,brand comparison,phone function,phone appearance,phone configuration and packaging gifts.Third,based on the analysis of emotional dictionary comment statements emotional tendencies,calculate each scored a comment of the emotion,and according to the emotional score will comment text is divided into positive comment,neutral and negative reviews,then use the LDA model,respectively,to extract the theme of the positive comments and negative comments,found that users of millet eight mobile phone price,brand,function,configuration and jingdong logistics speed is the positive evaluation,after-sales services and express delivery for jingdong mall,millet 8 the battery life of mobile phones and the Internet were negative fluency,also puts forward the requirements for mobile phone insurance service.Finally,according to the conclusion,the corresponding Suggestions are put forward for jingdong platform and xiaomi mobile phone suppliers.