This document discusses machine learning techniques for detecting credit card fraud. It begins with an abstract that outlines how credit card fraud causes major financial losses and how machine learning can help tackle this issue. It then provides background on credit card fraud and challenges in detecting it. The document describes the methodology used, including collecting transaction data, exploring relationships between features, and training models like random forests, decision trees, and support vector machines to classify transactions as fraudulent or legitimate. It finds these models achieved high accuracy scores between 99.7-99.8% but had low precision. The conclusion states that future work could focus on improving precision and considering additional algorithms and data processing techniques.