Deriving topic-related and interaction features to predict top attractive reviews for a specific business entity

ABSTRACT

As large volumes of online reviews are being generated, both online businesses and customers are confronted with big data challenges. Previous studies have developed various methods to predict the helpfulness of online reviews. These methods have disregarded the aspects of the business entities when dealing with datasets for prediction and evaluation and have not considered interactions between a review and the target business entity. In this paper, we propose a novel method to predict the top attractive reviews for a specific business entity. We also propose topic-related features to characterise the topics in a review and interaction features to reflect relationships between a review and the business entity it covers. Our empirical evaluation shows the utility of our proposed method and features.

KEYWORDS:

• Online review helpfulness
• top attractive reviews
• predictive analytics
• text mining
• machine learning
• latent topic model

Disclosure statement

No potential conflict of interest was reported by the authors.