Learning MultiLinguistic Knowledge for Opinion Analysis

详细信息    查看全文

• 作者：Ruifeng Xu ; Kam-Fai Wong ; Qin Lu ; Yunqing Xia
• 关键词：Opinion Analysis ; Linguistic Knowledge Learning ; Collocation
• 刊名：Lecture Notes in Computer Science
• 出版年：2008
• 出版时间：2008
• 年：2008
• 卷：5226
• 期：1
• 页码：993-1000
• 全文大小：202.7 KB

文摘

Most existing opinion analysis techniques used word-level sentiment knowledge but lack the learning capacity on the behaviors of context-dependent opinion words. Meanwhile, the use of collocation-level sentiment knowledge is not well studied. This paper presents an opinion analysis system, namely OA, which incorporates the word-level and collocation-level sentiment knowledge. Based on the observation on the NTCIR-6 opinion training corpus, some word-level and collocation-level linguistic clues for opinion analysis are discovered. Learning techniques are developed to learn the features corresponding to these discovered clues. These features are in turn incorporated into a classifier based on support vector machine to identify opinionated sentences and determine their polarities from running text. Evaluations on NTCIR-6 opinion testing dataset show that OA achieved promising overall performance.