Balanced random forest

This project is part of Algorithm Audit's knowledge base.

🌐 Knowledge base Algorithm Audit: website

🧰 Other tools and projects of Algorithm Audit: GitHub repo overview

Project Overview

This project contains the data pipeline for training and testing a balanced random forest (BRF) model. This model is applied on:

• Kaggle - Credit Card Fraud Detection, 284.807 transactions with 0.172% fraud rate. Goal of this repository is to elaborate what sensitivity testing can be performed to strike a balance between precision and recall

Methods used

• Balanced Random Forest (BRF)
• k-fold cross validation
• Precision-Recall
• Sensitivity testing