Fraudulent Amazon Reviewers

listen on castbox.fmlisten on google podcastslisten on player.fmlisten on pocketcastslisten on podcast addictlisten on tuninlisten on Amazon Musiclisten on Stitcher



Fraudulent Amazon Reviewers

On the show today, we are joined by Rajvardhan Oak, an applied Scientist at Microsoft. Raj works in the Ad Fraud division, and today, he discusses his personal research titled, The Fault in the Stars: Understanding the Underground Market of Amazon Reviews.

Raj began with a background on what he does working in the Ad Fraud department of Microsoft. He then shared an experience of how he got to know about fake reviews on Amazon. He explained the three kinds of fraud that exists, each from the buyers, sellers, and agent side. 

Raj delved deeper into how these fraudulent reviews work with real-life scenarios. He also discussed the efforts of Amazon to mitigate the activities of fraudulent reviewers. 

Raj then discussed the process of modeling the problem and finding the relevant data with ground truth. He spoke about the demographics of the agents and buyers in the fraudulent review web.

Raj discussed the quantitative and qualitative analysis he carried out on the dataset. He shared the discrepancy he observed between the control group and test group of the data. He also discussed the inferences that could be made from the discrepancy between both data groups.

Raj then spoke about the machine learning results in terms of precision and accuracy. He emphasized the need for further research for a more precise result at scale. Rounding up, Raj discussed how fraudulent review companies avoid being detected by Amazon. He finally discussed recommendations to forestall fraudulent reviews on e-commerce platforms. You can follow Raj on Twitter, Instagram, or LinkedIn.

Rajvardhan Oak

Rajvardhan Oak is a cyber security researcher passionate about making the Internet a safer place for everyone. As part of the ads fraud detection team at Microsoft, his work involves analyzing network traffic and building models to filter suspect fraud like click fraud, rewards fraud and competitive fraud. His work has helped keep the Microsoft ads network clean from fraud and recovered millions of dollars in revenue. He is also pursuing a PhD at the University of California Davis, where his research is focused on examining the underground ecosystems of reputation manipulation via fraudulent reviews. His investigation has uncovered the scale and nature of reviews fraud, and the operational characteristics of the ecosystem.

Thanks to our sponsors for their support

Astrato is a modern BI and analytics platform built for the Snowflake Data Cloud. A next-generation live query data visualization and analytics solution, empowering everyone to make live data decisions.
Log, store, query, display, organize, and compare all your model metadata in a single place