Predicting Helpfulness of Crowd-Sourced Reviews: A Survey

Abstract

With the recent advancement of Web 2.0 and the popularity of social media platforms, the volume of User Generated Content (UGC) is rising explosively. Online reviews are rapidly growing and a popular source of UGC, which help customers in evaluating the quality of product and making purchase decisions. However, distilling the required information from the massively increasing volume of reviews becomes difficult for customers. Therefore, it becomes an important issue to identify helpful review accurately. The review helpfulness prediction has attracted growing attention of researchers that proposed various solutions using statistical and Machine Learning (ML) techniques. This paper aims to review the existing literature on review helpfulness prediction, to identify data sources, ML techniques and potential challenges. The review helpfulness prediction was equally taken as both regression and classification task by previous studies. However, the definition of helpfulness for each task varies significantly. Most of the studies used online reviews from Amazon to predict helpfulness. The comparison of state-of-the-art techniques and challenges will give a quick overview to researchers about the existing state of research on review helpfulness prediction.