Department of Chinese, Translation and Linguistics
Research Degree Forum
Identifying Opinion Holders and Targets with Dependency Parser in Chinese News Texts
Mr. LU Bin
PhD candidate, Department of Chinese, Translation and Linguistics, City University of Hong Kong
Date: 5 Jul 2010, Monday
Time: 4:30 - 5:30 pm
Venue: B7603 (7/F, Blue Zone), Academic Building, CityU
Although there have been research on identifying opinion holders and targets in English product reviews and news texts, little work has been reported on the similar tasks on Chinese news texts. In this paper, we propose to identify opinion holders and targets in Chinese news texts by syntactic dependency structures.
Based on our linguistic analysis on opinions, holders and targets, we attempt to identify opinion holders by means of reporting verbs and to identify opinion targets by considering both the opinion holders and opinion-bearing words. The contributions from this study are: 1) we propose that the existence of reporting verbs is a very important feature for identifying opinion holders in news texts, which has not been clearly indicated; 2) we argue that the identification of opinion targets should not be done alone without considering opinion holders, because opinion holders are much easier to be identified in news texts and the identified holders are quite useful for the identification of the associated targets.
The experiments on NTCIR-7 MOAT’s traditional Chinese news test data show that our approach provides better performance on opinion holder/target identification than the baselines and most systems reported at NTCIR-7.
Mr. LU Bin is currently a PhD candidate from the Department of Chinese, Translation and Linguistics. His research interests includes Sentiment Analysis and Opinion Mining, Statistical Machine Translation (SMT), Computational Linguistics, and Natural Language Processing (NLP).
~ CTL Staff and Research Degree Students only ~