Generating unfair ratings to attack online rating systems

• Main Author: Neel Ratan Mitra
• Other Authors: School of Computer Engineering
• Format: Final Year Project
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Computer science and engineering
• Online Access: http://hdl.handle.net/10356/59075
• Institution: Nanyang Technological University

Trust plays an emphatic role in any form of transaction, the internet boom in the early 1990s brought in a new era of e-commerce allowing online transactions, reviews as some of the few activities out of many. These activities gave rise to countless rating systems to give consumers and sellers the opportunity to rate the quality of the transactions and the reputation of their counterpart to make others aware of the circumstances of a transaction they might be likely to get involved in. However, online rating systems also give malicious users the opportunity to generate unfair ratings, making the rating system flawed. Dealing with such attacks and threats have been an imperative yet challenging problem for a prolong period. This report focuses on the electronic simulated environment system developed to simulate an e-commerce market, the different attack and defense models based on single and multiple rating criteria. Real data has also been used to verify findings and ultimately it can be inferred that different evaluation metrics provide different perspectives in evaluating the efficiency of a trust model, these models also improve in their efficiency by not only accumulating ratings given to sellers and buyers over time but also by evaluating the trust of the advisors, who advice a user – buyer - on a common seller.