DocumentCode
1925956
Title
The Unified Collocation Framework for Opinion Mining
Author
Xia, Yun-qing ; Xu, Rui-Feng ; Wong, Kam-Fai ; Zheng, Fang
Author_Institution
Tsinghua Univ., Beijing
Volume
2
fYear
2007
fDate
19-22 Aug. 2007
Firstpage
844
Lastpage
850
Abstract
Opinion mining is a complicated text understanding technology involving opinion extraction and sentiment analysis. State-of-the-art techniques adopt idea of attribute-driven or sentiment-driven, leading to low opinion mining coverage. This paper proposes the unified collocation framework (UCF) and describes a novel unified collocation-driven (UCD) opinion mining method. The UCF incorporates attribute-sentiment collocations as well as their syntactical features to achieve reasonable generalization ability. Preliminary experiments show that 0.245 on averages improve recall of opinion extraction without obvious loss on opinion extraction precision and sentiment analysis accuracy.
Keywords
data mining; text analysis; attribute-sentiment collocation; low opinion mining coverage; opinion extraction; sentiment analysis; syntactical features; text understanding; unified collocation framework; unified collocation-driven opinion mining; Cybernetics; Data mining; Databases; Electronic mail; Machine learning; Manufacturing; Natural languages; Ontologies; Search engines; Speech analysis; Opinion extraction; Opinion mining; Sentiment analysis; Unified collocation;
fLanguage
English
Publisher
ieee
Conference_Titel
Machine Learning and Cybernetics, 2007 International Conference on
Conference_Location
Hong Kong
Print_ISBN
978-1-4244-0973-0
Electronic_ISBN
978-1-4244-0973-0
Type
conf
DOI
10.1109/ICMLC.2007.4370260
Filename
4370260
Link To Document