In the United States in 2013 alone, credit card fraud cost companies almost $7.1 billion dollars. Given these enormous costs, fraud detection and classification has become a very active area of research in machine learning and data mining domains. Although the power of machine learning techniques for fraud detection has greatly increased over the past decades, the incentives for fraudsters to circumvent and adapt to these classification algorithms has also grown. Effective fraud detection models must be able to adapt to behavioral changes on the part of the adversary, while maintaining high levels of accuracy and low levels of false positives. The credit card fraud team will investigate unsupervised methods with the goal of clustering customer types. The team will investigate how clusters can be used to improve fraud detection.
Banks and credit card companies provide many different cloud-based applications for their customers. When an application goes down, it can lead to significant financial losses for both the company and customers. Each of these applications generates data such as network traffic, resource utilization statistics and API calls which are all stored in AWS. This data can be joined to form time series data to make useful inferences about the state of each application. The main objective is to build an application health-monitoring system that predicts when an application crashes based on the time series data to minimize downtime. Secondary objectives include catching as many crashes as possible while minimizing false positives.