Building Chinese Sentiment Lexicon Based on HowNet

Abstract:

Judging the sentiment orientation of Chinese words is the basic work of the passage sentiment orientation research. Using Chinese basic sentiment words and corpus, we can identify sentiment words in the passage and expand sentiment lexicon effectively in order to improve the result of text semantic orientation analysis. With the basis of HowNet [1] sentiment words, we construct a Chinese sentiment lexicon by analyzing sentence structure and calculating the score of semantic similarity. We conduct Chinese text sentiment orientation classification experiment with this lexicon, the result shows the accuracy has achieved above 70% and obtained quite good classification effect.

• Periodical: Advanced Materials Research (Volume 187)
• Pages: 405-410
• DOI: https://doi.org/10.4028/www.scientific.net/AMR.187.405
• Online since: February 2011
• Authors: Yu Wei Liu, Shi Bin Xiao, Tao Wang, Shui Cai Shi
• Keywords: HowNet, Semantic Similarity, Sentence Structure, Sentiment Lexicon, Text Orientation Analysis

References

[1] Dong Z, Dong Q. HowNet, http: /www. keenage. com.

[2] Qun LIU , Sujian LI. Word Similarity Computing Based on How-net. CLSW3. (2002).

[3] Feng Li, Fang Li. A New Approach Measuring Semantic Similarity in Hownet 2000. Journal of Chinese Information Processing. (2007), 21(3) p.99.

[4] Jie Zhu, Gongshen Liu, Zhuo Chen. A Comparative Study on Semantic Orientation Classification of Chinese Text. Information Security and Communications Privacy. (2010).

[5] Songbo Tan. ChenSentiCorp. http: /www. searchforum. org. cn/tansongbo/corpus-senti. htm.

[6] Bin Wen, Ting-ting He, Le Song, Qian Wang. Text Sentiment Classification Research Based on Semantic Comprehension. Computer Science. (2010), 37(6).