Text Sentiment Analysis On E-commerce Review Data

With the popularization of the Internet,online shopping has become a necessity for people's lives.More and more categories have been involved in online shopping.The women's clothing industry has always been a giant development in e-commerce.For this reason,e-commerce reviews have become an important research topic.An important reference for consumers to choose the good in shopping is to read the previous collected consumer comments.Therefore,how to grab the useful information from collected consumers comments is very important.In other hand,how the merchants extract the consumers' concerns and complaints on their products from the consumers comments is helpful for the development of the companies.So as to design products that make consumers more satisfied and stand out in the increasingly fierce competition,it has become an important issue for e-commerce merchants.Taking Jingdong Mall Inman store's down jacket reviews as an example,we have built a sentiment classification model based on the collected corpus using the web crawler technology and the sentiment analysis technology.The specific steps are as following: Firstly,the scraped review data is preprocessed.Based on the python software Bayesian classification model,we obtained the sentimental tendency of the text reviews.In addition,we build a word cloud analysis chart based on above mentioned high-frequency words about the positive reviews and the negative reviews.By this way,the consumers' concerns and complaints about the product are obtained.Furthermore,the LDA topic model is further established to extract the potential topic words including in the reviews.It can clearly show the consumers' suggestions and concerns about the product.Based on above analysis,it is helpful for the merchants to make corresponding adjustments accordingly and to improve their competitiveness.