Please use this identifier to cite or link to this item:
Title: Learning multiLinguistic knowledge for opinion analysis
Authors: Xu, R
Wong, KF
Lu, Q 
Xia, Y
Keywords: Opinion analysis
Linguistic knowledge learning
Issue Date: 2008
Publisher: Springer
Source: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics), 2008, v. 5226, p. 993-1000 How to cite?
Journal: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics) 
Abstract: 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.
Description: 4th International Conference on Intelligent Computing, ICIC 2008, Advanced Intelligent Computing Theories and Applications. With Aspects of Theoretical and Methodological Issues, Shanghai, China, September 15-18, 2008
ISBN: 978-3-540-87440-9
ISSN: 0302-9743
EISSN: 1611-3349
DOI: 10.1007/978-3-540-87442-3_122
Appears in Collections:Conference Paper

View full-text via PolyU eLinks SFX Query
Show full item record

Page view(s)

Last Week
Last month
Citations as of Jul 10, 2018

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.