Machine Learning Applications in E-Commerce Fraud Detection

Details
Internet fraud has costed billions of dollars in the past decade. For websites with medium to large user traffic, it is important to leverage various technologies to manage the exposure and risk from online fraud. Machine Learning technology and products can play a critical role in building a strong defense line against online fraud. In this talk, Tong will discuss some of the best practices in this application field. The talk will include discussions on ML algorithms like logistic regression and Neutral Networks, and also include broader discussions on how ML fits into the overall picture of fraud detection and non ML-based solutions that combat fraud effectively.
Speaker Bio: Tong Li joined eBay in 2002 as one of the founding team members of eBay Trust and Safety group. He built the first machine learning fraud detection model at eBay, led a 50+ people Trust Science team at eBay in the past decade, and worked with partner teams to build eBay fraud detection system from ground up to protect billions of transactions and hundreds of millions of users from global online fraudsters. Prior to eBay, Tong served as the head of underwriting modeling team in Providian focusing on credit and fraud risk management in financial service industry. Tong is currently working on getting an early stage data science startup off the ground.

Machine Learning Applications in E-Commerce Fraud Detection