作者: Xiang-Wen Liao , Hu Chen , Jing-Jing Wei , Zhi-Yong Yu , Guo-Long Chen
DOI: 10.1109/ICMLC.2014.7009715
关键词:
摘要: In recent years, opinion retrieval attracted a growing research interest as online users' opinions become more and valuable for market survey, political polls, etc. The goal of is to find relevant opinionate documents according user's query. Compared with previous lexicon-based generative model considering that the sentiment words are equal query, which cannot reflect different words' strength, we propose graph-based approach by using HITS capture strength. Then weights incorporated into weighted retrieval. Experimental results on two datasets show effectiveness proposed model. baseline approach, improvements 4% 11% have been obtained real datasets.