An Improved Method of Sentiment Analysis of Chinese Web Reviews

Author

Jianzhuo Yan ; Pengying Li ; Liying Fang

Author_Institution

Sch. of Electron. Inf. & Control Eng., Beijing Univ. of Technol., Beijing, China

Volume

1

fYear

2014

Firstpage

351

Lastpage

355

Abstract

The web reviews contain a lot of features of the evaluation object and express the viewpoints or attitudes of the reviewers. Therefore, analyzing them and identifying the sentiment polarity will be significantly important and valuable. This paper uses the appearing-frequency of noun words to extract the aspect words and uses some semantic rules to extract opinion words to form the aspect-opinion word pairs. We use the sentiment orientation-point mutual information (SO-PMI) algorithm to calculate the sentiment strength of opinion words and get the semantic orientation of the opinion words and improved the algorithm. With the Chinese reviews of doctors from the medical websites www.39.net and www.chan120.net, experimental results demonstrate the validity of the proposed method.

Keywords

Internet; natural language processing; Chinese Web reviews; SO-PMI algorithm; appearing-frequency; aspect-opinion word pairs; medical Web sites; noun words; opinion word extraction; opinion words; semantic orientation; semantic rules; sentiment analysis; sentiment orientation-point mutual information algorithm; sentiment polarity; sentiment strength; www.39.net; www.chan120.net; Algorithm design and analysis; Compounds; Feature extraction; Medical services; Semantics; Sentiment analysis; Support vector machines; aspect extraction; opinion mining; sentiment analysis; sentiment strength;

fLanguage

English

Publisher

ieee

Conference_Titel

Computational Intelligence and Design (ISCID), 2014 Seventh International Symposium on

Print_ISBN

978-1-4799-7004-9

Type

conf

DOI

10.1109/ISCID.2014.237

Filename

7064208