Portfolio 3: End-to-End Credit Risk Modelling & Portfolio Risk Analytics (UK)
An end-to-end credit risk modelling project using a large-scale synthetic UK lending dataset to estimate Probability of Default (PD) and quantify portfolio risk through LGD, EAD, and Expected Loss (EL) under stressed macroeconomic conditions. The framework combines interpretable Logistic Regression with advanced machine learning models, enhanced by LIME for local explainability, to support risk-based pricing, optimise decision thresholds, and enable data-driven, Basel II/III-aligned credit decision-making. 
.png)
.png)