Classification Research Of Sentiment Analysis For Product Reviews

Into the twenty-first century, with the rapid development of internettechnology and e-commerce area, there are more and more products reviews in thenetwork. These reviews have become very important source from which customersobtain information. But it is too hard to read for people because of increasingreviews. How to confirm the appraise of products reviews correctly has become aurgent problem.Analysis on appraise of products reviews reflects customers’ satisfaction offunction and provide sellers timely feedback information. It is an importantapplication area in artificial intelligence and sentiment analysis. At present,customers’ review texts mainly have a study of evaluate sentiment score and rarelyconsider the influence from internet catchphrases and complicated sentenses. Thispaper use two methods to analyze the sentiment orientation on a basis of steps ofsentiment analysis.At the beginning, this paper introduces the research background, includingdevelopment of internet, increased product reviews and the desire that people wantto confirm sentiment orientation in reviews. Secondly, it describes sentimentanalysis status at home and abroad. Thirdly, this paper introduces two methodsbased on text classification and sentiment dictionary for sentiment anylysisaccording to the description of gathering and preprocessing product reviews, thenwe use these two proposed methods to analyze the sentiment orientation of data.Finally, it summarize and outlook the sentiment analysis work.