Source Code for "A Reproducible Evaluation Framework for Benchmarking Machine Learning and Hybrid Ensemble Models in Power System Anomaly Detection"

Résumé fourni par la source

Description This repository contains the complete implementation of the study entitled "A Reproducible Evaluation Framework for Benchmarking Machine Learning and Hybrid Ensemble Models in Power System Anomaly Detection." The repository provides the full Python source code and Jupyter Notebook required to reproduce all experiments reported in the manuscript. The implementation includes: Data preprocessing pipeline Feature engineering and integrity verification Stratified train/validation/test partitioning Repeated Stratified 5-Fold Cross-Validation (5 × 10 repetitions) Random Forest Logistic Regression Support Vector Machine (RBF) K-Nearest Neighbors Gradient Boosting Voting Hybrid Ensemble Stacking Hybrid Ensemble Performance evaluation Confusion matrices ROC analysis SHAP explainability LIME explainability Noise robustness analysis Automatic generation of all tables and figures reported in the manuscript Reproducibility outputs including exact prediction files, confusion matrices, and evaluation metadata. The implementation was developed to ensure complete transparency and reproducibility of the experimental results presented in the accompanying manuscript.

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