Research On The Key Technologies Of Chinese Online Product Review’s Sentiment Analysis

The rapid development of e-commerce makes the online product reviews grow rapidly, which makes people can’t be easy to optimize their purchase decisions. As we all know, it’s difficult to get the valuable information from a flood of online reviews. So it’s necessary to alleviate the burden of reading through certain technical means which mainly refer to the sentiment analysis of massive online product reviews. At present, researchers have made some achievements in English. But the research of sentiment analysis about Chinese product comments is relatively few. This paper takes the Chinese online product reviews as research object. It mainly focuses on three key areas:the automatic product features extracting method, the method to build a Chinese sentiment lexicon and the sentiment analysis technology about sentiment summarizer.Firstly, we propose a novel unsupervised approach to extract product features from Chinese customer reviews. The extraction system firstly extracts the valuable seeds automatically with the effective pruning method. Then new features are extracted through some language rules. In order to assure the recall of our method, the feature weight is computed to dig more features. Our experiment results show that the algorithm gained good and stable performance in different domains.Secondly, this paper presents a method to automatically generate and score a sentiment lexicon for Chinese sentiment analysis. Node represents word and weighed edge stands for the relationship between two words in a weight graph. The polarity detection is achieved through the label propagation in this weight graph. Moreover, the lexicon is expanded through some direct relations between the words in a thesaurus. The experimental results demonstrate that the proposed algorithm can effectively improve the accuracy of the approach for constructing a sentiment lexicon.At last, based on the above results, we research the comprehensive information mining refer to the product features and the sentiment analysis based on product features. Meanwhile, the histogram is used to show the mining results (sentiment summarizer). Here, two tasks are performed, identifying product features and opinion orientations from each review which are based on our lexicon. From the histogram, people are able to see the strengths and weaknesses of each product in the minds of consumers clearly.In conclusion, this paper proposes some new algorithms to solve some key problems of Chinese online product review’s sentiment analysis. These studies will hopefully help users access to the product feedback information easily. Therefore, these researches will make the application of Chinese review mining technologies more widely in the field of e-commerce.