Data package for "Deterministic Machine-Learning Post-Processing of Day-Ahead Electricity Price Benchmarks: Direct and Residual Correction under Market Stress Conditions"

This repository contains the processed modeling dataset, final test-period predictions, numerical result workbooks, supporting tables, figures, and run manifest for the manuscript “Deterministic Machine-Learning Post-Processing of Day-Ahead Electricity Price Benchmarks: Direct and Residual Correction under Market Stress Conditions.” The raw source data are publicly available from the Kaggle dataset “Hourly Energy Demand, Generation and Weather” by N. Camattari. The files deposited here support the validation, final test, regime-aware, temporal robustness, sign-correction, rolling-origin, seed-sensitivity, residual autocorrelation, and asymmetric-risk results reported in the manuscript.