README.md

metadata

license: mit
tags:
  - fintech
  - reconciliation
  - classification

FinTech Reconciliation Break Classifier (Pair-Level)

This repository contains a trained model that predicts whether a bank↔broker reconciliation pair is a TRUE_BREAK (1) or FALSE_BREAK (0).

Features Used

['large_amt_mismatch', 'amt_diff', 'amt_pct', 'amt_diff_abs', 'amt_ratio', 'settlement_gap_days', 'is_weekend', 'is_settlement_weekend']

Best Run (selected by PR-AUC)

• Experiment: exp_001_baseline
• PR-AUC: 1.000
• ROC-AUC: 1.000
• F1: 0.998

Files

• model.joblib : trained sklearn pipeline
• features.json: selected feature list
• config.json : training metadata + best metrics
• predict.py : minimal inference helper