Article Preview
Top1. Introduction
Review helpfulness is the part of business intelligence (BI) and plays a pivotal role for the e-commerce business to populate their sites with number of genuine reviews to assist customers by their products and services. Google provides cumulative rating of a product based on the reviews and rating that it receives from various sources like pbtech, eBay, and Samsung. Online customers mainly like to read the online reviews related to that particular product, which they want to buy. Nowadays, plenty of reviews are available to help the online customers in deciding them about the right product. These days’ e-commerce websites try to discern the usefulness of reviews by conducting an online survey or through telephonically. Online review sites such as Yelp, Amazon, etc., offers millions of customer reviews, which might have greater impact on the market trends as well as on buying decisions of many potential customers (Guo et al., 2020). Aa per Murphy et al., 2020, about 86% of customers used to read online reviews and shows a deeper trust in them. By year 2020, more than 200 million reviews were published on Yelp. The review helpfulness provides the insight about the subjectivity and quality of the product (Li et al., 2019; Filieri et al., 2018). In general, helpful votes are treated as a true indicator of review reliability (Huang et al., 2015). The prime challenges of customers and businesses firms are to get the benefit from such reviews which are available in bulk quantity and are of inconsistent nature. Out of the several aspects of the reviews, helpfulness is taken as key feature and researched widely. Review helpfulness is computed by calculating the number of helpful votes divided by the total number of votes (Bilal et al., 2021). Figure 1 shows how Amazon.com gathers helpful votes of the reviews from their readers (Park et al., 2018). A lot of work in the direction of predicting helpfulness has already been carried out as of now, details of which are given in the subsequent section.
Figure 1. A typical review on Amazon (Source: Amazon, 2022)
Top2. Literature Review
Online product reviews are suggested as a valuable tool for promoting products, as well as for collecting consumer feedback and boosting sales (Chua et al., 2014; Forman et al., 2008; Hu et al., 2008). In literature (Gang et al., 2008), there have been shown the direct relationship between product ratings and sales. For example, online movie reviews and ratings have significant impact on box office revenues (Lee et al., 2018). Similarly, online book reviews have positive impact on book sales (Chevalier et al., 2006). Authors in (Zhang et al., 2006), put forward a regression model to predict the utility of product reviews. In these studies, the authors utilized lexical similarity, and syntactic terms based on Part-of-Speech (POS) as features.