Machine Learning Applications in Risk Management: Fraud Detection Using Machine Learning

Credit card fraud may be one of the most common fraudulent activities in many countries. However, the number of fraudulent activities is very small (less than 1%). Common performance metrics, such as accuracy, may not be that useful for determining model performance. In this demo, you will learn how to use machine learning to detect fraudulent activities as well as how to use built-in functions in MATLAB^® to calculate the area under the precision-recall curve (AUPRC), a custom performance metric.