Train the Fraud Classification Model
Builds a Random Forest classification model from the prepared training data. The model learns patterns linked to fraud vs. non-fraud transactions and produces both the trained model and training-time diagnostics, which are then used for later prediction on unseen data.
