LAUSR

Since Internet has become an excellent source of consumer reviews, the area of sentiment analysis (also called sentiment extraction, opinion mining, opinion extraction, and sentiment mining) has seen a large increase in academic interest over the last few years. Sentiment analysis mines opinions at word, sentence, and document levels, and gives sentiment polarities and strengths of articles. As known, the opinions of consumers are expressed in sentiment Chinese phrases. But due to the fuzziness of Chinese characters, traditional machine learning techniques can not represent the opinion of articles very well. In this paper, we propose a multi-strategy sentiment analysis method with semantic fuzziness to solve the problem. The results show that this hybrid sentiment analysis method can achieve a good level of effectiveness.