End-to-End Python implementation of Regime-Weighted Conformal (RWC) prediction for sequential VaR control in nonstationary financial markets (Schmitt, 2026). Combines kernel-based regime similarity with exponential time decay to calibrate distribution-free risk bounds. CRSP data validation, GBDT quantile forecasting, and rigorous backtesting.
-
Updated
Feb 8, 2026 - Jupyter Notebook