Key Issues on Feature-based Opinion Mining Study

The Internet contains vast amounts of subjective information. Feature-based opinion mining uses natural language processing, information extraction and data mining techniques to identify and analyze subjective information On the features of the object as well as the emotional tendencies. In this paper, we mainly introduce the basic concepts of feature-based opinion mining. Also the research ideas and techniques to be used are introduced. Then we describe the key issues of feature-based sentiment analysis and opinion mining. These key issues include: feature extraction, the calculation of features' weight, opinion words extraction, opinion holder extraction and sentiment classification, sentiment comparative sentences and negative sentences sentence processing, adverbs processing as well as Corpus construction.