Capstone project for Udacity's Machine Learning Engineer Nanodegree
The goal is to train a suitable model that can generalize well enough to predicting fraudulent transactions given the least amount of input variables possible.
The dataset can be downloaded from Synthetic data from a financial payment system Dataset
All development Python code used this your project is available as a Jupyter Notebook in fraud_detection.ipynb
A final report explaining this project and the surrounding problem domain is available as report.pdf
.