Please use this identifier to cite or link to this item: http://hdl.handle.net/10625/49051
Title: Structure-Aware Review Mining and Summarization
Authors: Li, Fangtao
Han, Chao
Huang, Minlie
Zhu, Xiaoyan
Xia, Ying-Ju
Keywords: LINGUISTIC RESEARCH
CLUSTER ANALYSIS
DATA MINING
MACHINE LEARNING
Date: Aug-2010
Citation: Li, F., Han, C., Huang, M., Zhu, X., & Xia, Y. (2010). Structure-Aware Review Mining and Summarization. Proceedings of the 23rd International Conference on Computational Linguistics (Coling 2010), Beijing, CHN. (p. 653-661).
Abstract: In this paper, we focus on object feature 1 1 Introduction based review summarization. Different from most of previous work with linguistic rules or statistical methods, we formulate the review mining task as a joint structure tagging problem. We propose a new machine learning framework based on Conditional Random Fields (CRFs). It can employ rich features to jointly extract positive opinions, negative opinions and object features for review sentences. The linguistic structure can be naturally integrated into model representation. Besides linear- chain structure, we also investigate conjunction structure and syntactic tree structure in this framework. Through extensive experiments on movie review and product review data sets, we show that structure-aware models outperform many state-of-the-art approaches to review mining.
Description: Proceedings of the 23rd International Conference on Computational Linguistics(Coling 2010)
URI: http://hdl.handle.net/10625/49051
Project Number: 104519
Project Title: International Research Chairs Initiative (IRCI)
Access: IDRC Only
Access Restriction: Due to copyright restrictions the full text of this research output is not available in the IDRC Digital Library or by request from the IDRC Library. / Compte tenu des restrictions relatives au droit d`auteur, le texte intégral de cet extrant de recherche n`est pas accessible dans la Bibliothèque numérique du CRDI, et il n`est pas possible d`en faire la demande à la Bibliothèque du CRDI.
Files in This Item:
There are no files associated with this item.
Show full item record