Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6854071 | Electronic Commerce Research and Applications | 2018 | 10 Pages |
Abstract
Identifying helpful reviews from massive review data has been a hot topic in the past decade. While existing research on review helpfulness estimation and prediction is primarily sourced from English reviews, non-English reviews may also provide useful consumer opinion information and should not be neglected. In this study, we propose a review helpfulness prediction framework that processes and uses multilingual sources of reviews to generate relevant business insights. Adopting a design science research approach, we design, implement, evaluate and deliver an IT artifact (i.e., our framework) that predicts the helpfulness of a review and accounts for non-English reviews. Our evaluations suggest that we achieve better performance on review helpfulness prediction and classification by including the variables generated by our instantiated multilingual system. By demonstrating the feasibility of our proposed framework for multilingual business intelligence applications, we contribute to the literature on business intelligence and provide important practical implications to practitioners.
Related Topics
Physical Sciences and Engineering
Computer Science
Artificial Intelligence
Authors
Ying Zhang, Zhijie Lin,