Source Code for "Explainable Machine Learning Framework for Dynamic Line Rating Forecasting with Adversarial Robustness Evaluation"

Résumé fourni par la source

This repository contains the source code associated with the manuscript entitled “Explainable Machine Learning Framework for Dynamic Line Rating Forecasting with Adversarial Robustness Evaluation”. The provided implementation supports the computational workflow used in the study, including data preprocessing, temporal feature engineering, LightGBM- and CatBoost-based DLR forecasting, model performance evaluation, adversarial robustness assessment using FGSM and BIM attacks, and explainability analyses using SHAP and LIME. The repository also contains the code used for clean-versus-adversarial SHAP analysis and for reproducing the corresponding feature-importance comparison presented in Figure 16. The source code is archived to support transparency, reproducibility, and independent verification of the computational analyses reported in the associated manuscript.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.