Code for: Machine Learning-Based Prediction of Wind and Solar Curtailment in Northeast Brazil

Python scripts for very short-term (30-minute ahead) forecasting of renewable energy curtailment in the Northeast Brazilian subsystem using ensemble machine learning algorithms. This repository contains:- Model comparison scripts (LightGBM, XGBoost, Random Forest)- Rolling window cross-validation analysis- Temporal and magnitude-based error analysis- Persistence benchmark calculation The code accompanies the paper: "Machine Learning-Based Prediction of Wind and Solar Curtailment: A Case Study of Northeast Brazil" Data source: Brazilian National System Operator (ONS) - https://dados.ons.org.br/