Lexicon knowledge extraction with sentiment polarity computation

• Main Authors: WANG, Zhaoxia, TONG, Vincent Joo Chuan, RUAN, Pingcheng, LI, Fang
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2016
• Subjects: Weibo; Chinese microblog; Domain knowledge building; Lexicon knowledge extraction; Natural Language Processing; Sentiment analysis; Artificial Intelligence and Robotics; Numerical Analysis and Scientific Computing
• Online Access: https://ink.library.smu.edu.sg/sis_research/5488; https://ink.library.smu.edu.sg/context/sis_research/article/6491/viewcontent/Lexicon_av.pdf
• Institution: Singapore Management University

Sentiment analysis is one of the most popular natural language processing techniques. It aims to identify the sentiment polarity (positive, negative, neutral or mixed) within a given text. The proper lexicon knowledge is very important for the lexicon-based sentiment analysis methods since they hinge on using the polarity of the lexical item to determine a text's sentiment polarity. However, it is quite common that some lexical items appear positive in the text of one domain but appear negative in another. In this paper, we propose an innovative knowledge building algorithm to extract sentiment lexicon knowledge through computing their polarity value based on their polarity distribution in text dataset, such as in a set of domain specific reviews. The proposed algorithm was tested by a set of domain microblogs. The results demonstrate the effectiveness of the proposed method. The proposed lexicon knowledge extraction method can enhance the performance of knowledge based sentiment analysis.