期刊名称：Humanities and Social Sciences Communications
电子版ISSN：2662-9992
出版年度：2020
卷号：7
期号：1
页码：1-12
DOI：10.1057/s41599-020-00578-9
语种：English
出版社：Springer
摘要：While electronic word-of-mouth (eWOM) variables, such as volume and valence have been posited in previous studies to consistently affect product sales, there is a lack of studies on the different contexts and outcomes that affect the importance of eWOM variables. In order to fill this gap, this study attempts to use the helpfulness of reviews and reviewers as moderators to predict box office revenue, comparing the prediction performances of business intelligence (BI) methods (random forest, decision trees using boosting, the k-nearest neighbor method, discriminant analysis) using eWOM between high and low review or reviewer helpfulness subsample in the Korean movie market scrawled from the Naver Movies website. The results of applying machine learning methods show that movies with more helpful reviews or those that are reviewed by more helpful reviewers show greater prediction performance, and review and reviewer helpfulness improve the prediction power of eWOM for box office revenue. The prediction performance will improve if the characteristics of eWOM are likely to be combined to contribute to box office revenue to a greater extent.
