README.md

Credit-Risk-Modeling# Credit Risk Modeling

Four projects covering the lending lifecycle from underwriting through impairment accounting:

1. PD Modeling — the backbone: probability the borrower defaults in 12 months. Logistic regression baseline + XGBoost challenger.
2. LGD Modeling — how much of a defaulted loan you actually lose, after recoveries. Two-stage cure + severity model.
3. Credit Scoring — application scorecard scaled to FICO-style points. The format underwriters and adverse-action notices actually need.
4. IFRS 9 ECL — wires PD × LGD × EAD together with stage classification to produce the loan-loss provision that lands in the income statement.

Each subfolder has its own README walking through the methodology, validation results, and what would need to change for a real production deployment.

Suggested reading order

If you're new to the field, work through them in this order:

1. pd-modeling/ (the model itself)
2. credit-scoring/ (how the model gets deployed in underwriting)
3. lgd-modeling/ (the second leg of the loss equation)
4. ifrs9-ecl/ (how it all rolls up into accounting numbers)