Research on Building a Chinese Sentiment Lexicon Based on SO-PMI

Abstract:

Considering user behavior, this paper has built a Chinese sentiment lexicon based on improved SO-PMI algorithm. Sematic lexicons were used to classify the sentiment of the collected Chinese hotel reviews. The experiment has compared the feature extraction between CHI and sentiment lexicons to find out different classification performances. The results indicate that feature extraction based on sentiment lexicon gains higher F₁. The performance of classification method “Basic Semantic Lexicon + BOOL + NB” gains 92.40% of F₁. Based on different sentiment lexicons, the experimental results shows that (SO-A) and (SO-P) is slightly better than NB classifier. Therefore, it would be effective to use ((SO-A) and (SO-P) as text sentiment classifiers. The experiment also finds out the method “Hotel Reviews Semantic Lexicon using improved SO-PMI algorithm +((SO-A)” gains the highest F₁ which is 92.84%. The results reveal that improved SO-PMI does more effective on weight calculation and sentiment lexicon building.

• Periodical: Applied Mechanics and Materials (Volumes 263-266)
• Pages: 1688-1693
• DOI: https://doi.org/10.4028/www.scientific.net/AMM.263-266.1688
• Online since: December 2012
• Authors: Ai Min Yang*, Jiang Hao Lin, Yong Mei Zhou, Jin Chen
• Keywords: Improved SO-PMI, Naive Bayesian, Sentiment Classification, Sentiment Lexicon

References

[1] Ding Yang, Aimin Yang. Classification approach of Chinese texts sentiment based on semantic lexicon and naive Bayesian[J]. Application Research of Computers, 2010,27(10):3737-3739. In Chinese.

[2] Liuling Dai etc.Measuring Semantic Similarity between Words Using HowNet[J]. ICCSIT2008, 2008: 601-605.

[3] Weiping Liu etc.Research on building Chinese basic semantic lexicon[J].Journal of Computer Applications,2009,29(11):2882-2884. In Chinese.

[4] Yanhui Zhu etc. A Method of Emotional Feature Extraction in Chinese Text Based on Multiple Lexicons[J] Journal of Hunan University of Technology,2011,25(2):42-46. In Chinese.

[5] Stefano Baccianella, Andrea Esuli, and Fabrizio Sebastiani.SENTIWORDNET3.0: An Enhanced Lexical Resource for Sentiment Analysis and Opinion Mining[J].LREC,2010:2200-2204

[6] Alaa Hamouda, Mahmoud Marei and Mohamed Rohaim.Building Machine Learning Based Senti-word Lexicon for Sentiment Analysis[J].Journal of advances in information technology, 2011,2(4):199-203.

DOI: 10.4304/jait.2.4.199-203

[7] Ronglu Li. Research on Text Categorization and Technologies [D].Shanghai: Fudan University, 2005. In Chinese.

[8] TURNEY P D, LITTMAN M L. Measuring Praise and Criticism: Inference of Semantic Orientation from Association[J]. ACM Transactions on Information Systems. 2003,21(4): 315-346.

DOI: 10.1145/944012.944013

[9] LCWCC Workgroup: Lexicon on Common Words of Contemporary Chinese [K]. Beijing: The Commercial Press,2008. In Chinese.

[10] Jianghao Lin et. Classification of microblog sentiment based on naïve Bayesian[J]. Computer Engineering & Science,2012,34(9):160-165. In Chinese.