Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4972643 | Information & Management | 2017 | 41 Pages |
Abstract
Motivated by the role of product features in enabling personalized recommendations and marketing, this research aims to extract product features from online consumer reviews. Previous studies are dominated by statistical-based techniques or focused on subjective features that are associated with opinions. In this research, we propose RubE-unsupervised rule-based methods that extract both subjective and objective features from online consumer reviews. We identify objective features by incorporating part-whole relation and review-specific patterns. We extract subjective feature by extending double propagation with indirect dependency and comparative construction. The experiment results demonstrate that RubE significantly outperforms the state-of-the-art techniques for product feature extraction and is generalizable from search goods to experience goods.
Related Topics
Physical Sciences and Engineering
Computer Science
Information Systems
Authors
Yin Kang, Lina Zhou,