From Positive Feedback to Comprehensive Rating: An Auto-Rating Models for Online Fictions in Sharing Communities

Abstract

Generally, 'rating' is a classification that represents a product's quality, satisfaction, or performance. Ratings are extensively used for ranking, which helps consumers decide to purchase or use the product. However, free products in sharing communities, such as artworks, comics, and fiction, often receive only positive or neutral feedback. This limitation hampers the rating and ranking capabilities of such products in recommending them to customers. Consequently, our research focuses on the rating of online fiction, which is a growing trend in sharing communities and is widely available on various online platforms, including websites and mobile applications. This research aims to develop a generic rating model for automated rating generation. We propose integrated rating models built by extracting factors from online platforms and social media derived from fiction readers' communities. The ratings generated by our proposed models are compared with the top 10 favourite Boylove fiction rankings from ReadAwrite, a Thai online fiction platform, using Pearson's correlation coefficient. The results shed light on the potential implementation of an auto-generation rating system for free products.