Identifying Fraudulent E-Commerce Behaviour via Robust Anomaly Detection

The objective is to develop a real-time fraud detection algorithm for a large E-commerce company based on construction of a robust reference model from normal multivariate data. To accomplish this, we propose to leverage machine learning techniques, such as reinforcement learning, in combination with stochastic modeling techniques such as Hidden Markov Models, to provide both a comparative study between the approaches and possibly produce an enhanced algorithm which applies both methods appropriately. The intern will be exposed to a variety of new methodologies, acquire a deep understanding of techniques in operations research and its application to fraud prevention.

Faculty Supervisor:

Viliam Makis

Student:

Larkin Liu

Partner:

Paytm Labs

Discipline:

Engineering - mechanical

Sector:

University:

University of Toronto

Program: