GitHub - Jahetbe/Terrain-Driven-Wildfire-Forecasting-with-Explainable-AI-A-Case-Study-from-Alberta: This study develops an interpretable wildfire susceptibility framework for Alberta’s boreal forests using ensemble ML models (LightGBM, XGBoost, RF, GB, AdaBoost) with exhaustive feature selection and Bayesian optimization. Among 65,451 configurations, LightGBM achieved >98% accuracy using 8 key predictors.

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
All_Input_Combinations_of_the_best_Method		All_Input_Combinations_of_the_best_Method
Ensemble_Models		Ensemble_Models
Ensemble_Optimized		Ensemble_Optimized
SHAPs_Spearman_VIF		SHAPs_Spearman_VIF

About

This study develops an interpretable wildfire susceptibility framework for Alberta’s boreal forests using ensemble ML models (LightGBM, XGBoost, RF, GB, AdaBoost) with exhaustive feature selection and Bayesian optimization. Among 65,451 configurations, LightGBM achieved >98% accuracy using 8 key predictors.