Combining Domain-Specific Sentiment Lexicon with Hownet for Chinese Sentiment Analysis

Various sentiment analysis approaches have been proposed but little has been done for Chinese documents. In this paper, a small set of Chinese adjective lexicon has been constructed in which each word is context-independent and is labeled with polarity intensity manually. Meanwhile, we introduce a framework which combines context-sensitive sentiment lexicon and the existing sentiment word list such as “Hownet” to enhance the domain specific opinion data analyzing performance. An evaluation experiment which uses the real online product reviews as input also discussed in this paper.