基于句法分析的商品评论特征分析

With the development of E-Commerce web 2.0 promoted by technology of web 2.0, people do shopping on the net and put their comments on the products.Regarded as feedback information of user's experience, product reviews is becoming more and more important and how to comprehend these reviews become an important research topic in web mining. In this paper, an product reviews mining based on semantics comprehends and feature anlysis is introdued. First，Tag POS for the word of the sentence，extract the feature of the product and calculate the weight of adjective and adverbs by TFIDF。After that，Modify the polarity according to the relative position between subject and comparative word and semantic character of the comparative word，base on the method of SO-PMI. The experimental results show that the algorithm effectively estimate the diversity of feature among these product and improve accuracy.