Cross-Domain Sentiment Analysis of Product Reviews by Combining Lexicon-based and Learn-based Techniques

Product reviews can direct consumers' purchasing behaviors and sellers' marketing strategies. Therefore, in this paper, we propose a novel sentiment analysis method which Combines Lexicon-based and Learn-based techniques (CLL) to analyze the cross-domain sentiment of Chinese product reviews. We first build three domain lexicons based on the basic lexicon and corpus from three domains containing books, hotels and electronics. Furthermore, we use four categories of features (including 16 features in total) to build six classifiers. We conduct a series of experiments to evaluate our proposed CLL by using different lexicons and different classifiers. Our experimental results show that domain lexicons outperform the basic lexicon no matter in which domain. Our method CLL performs better than state-of-the-art methods in domains of books and hotels, and is slightly inferior in the domain of electronics.