Feature-based Sentiment Analysis Approach for Product Reviews

The researches and applications of sentiment analysis become increasingly important with the rapid growth of online reviews. But traditional sentiment analysis models have been lacking in concern on the modifying relationship between words for sentiment analysis of Chinese reviews, and limit the development of opinion mining. This paper proposes a feature-based vector model and a novel weighting algorithm for sentiment analysis of Chinese product reviews. It considers both modifying relationships between words and punctuations in review texts. Specifically, it can classify reviews into two categories, i.e., positive and negative, and can also represent the sentiment strength by adverb of degree. Moreover, a novel feature extraction method based on dependency parsing is presented to identify the corresponding aspects that opinions words modify. We conduct some experiments to evaluate our algorithms, and demonstrate that the proposed approaches are efficient and promising.