Wildfire Susceptibility Mapping Using Five Boosting Machine Learning Algorithms: The Case Study of the Mediterranean Region of Turkey

Table 5

The prediction performance results of the models according to the measurement methods.

Measurements	LightGBM		GBM		XGBoost		AdaBoost		CatBoost
	Train	Test	Train	Test	Train	Test	Train	Test	Train	Test
Overall accuracy	99.886	94.703	85.675	84.488	90.443	88.801	86.430	86.021	99.126	95.473
Precision	0.999	0.933	0.822	0.816	0.871	0.863	0.840	0.855	0.989	0.951
Recall	0.999	0.939	0.844	0.838	0.884	0.885	0.854	0.873	0.990	0.954
Sensitivity	0.999	0.956	0.873	0.853	0.929	0.892	0.876	0.846	0.993	0.956
Specificity	0.999	0.939	0.844	0.838	0.884	0.885	0.854	0.873	0.990	0.954
AUC	0.999	0.948	0.858	0.846	0.907	0.888	0.865	0.859	0.991	0.955
F1	0.999	0.936	0.833	0.827	0.877	0.874	0.847	0.864	0.990	0.952
Kappa Index	0.998	0.894	0.712	0.689	0.808	0.774	0.728	0.719	0.982	0.909