AmazonRep: A Reputation System to Support Amazon’s Customers Purchase Decision Making Process based on Mining Product Reviews

摘要

In this paper, we develop AmazonRep, a reputation system to support Amazon’s customers purchase decision making process based on mining product reviews and their attributes. Different from existing reputation systems that incorporate review sentiment orientation, review helpfulness votes and review time for the purpose of generating and visualizing reputation toward various e-commerce items (movies, products, hotels, services…), disregarding important factors such as user credibility, our system mainly focus on considering all these factors (review sentiment orientation, review helpfulness votes, review time and user credibility) during reputation generation and visualization. Experimental results coming from an analysis of 13 Amazon’s products (books, movies, laptops, smartphones, video games, washing machines and refrigerators) demonstrate the usefulness of the proposed system in generating and visualizing reputation.