This script loads several financial datasets, cleans and merges them into a single training table, performs feature engineering, and builds a machine‑learning model to predict whether an account will have long payment delays. It trains 2 models a standard Random Forest and a Differential Privacy Random Forest & compares their accuracy.