Computing Early Reviewers Using Spam Filtration for Effective Product Marketing on E-Commerce Websites

Abstract

Now-a-days online reviews have become a significant basis of information for customers before they take any purchase decision. Most of the buyers trust the online reviews before buying any product online. Early review is the review given at the initial stage of the product life cycle and it has a very significant effect on the product sale. In this paper, basic behavioral characteristics of the reviewers who have given the early reviews have been analyzed based on giant e-commerce website, Amazon. Early reviewer is the user who gives a review on any product in its early stage. Early reviewers are important because of their rating activities, the helpfulness score that received by them from other users and the association of their reviews with the quality of product. Spam reviews are those reviews which are potential to affect the product quality as well as the product business. Spam reviews are falsifying, partially true or irrelevant to the product. Unauthenticated reviews, redundant reviews or null reviews are addressed as the spam reviews. Spam reviews are addressed as a part of data preprocessing and early reviewers are calculated using Random Forest Algorithm.