Table 3.

Results of each hierarchical model.

Preprocessing and Feature Selection Methods	INFO-SVM				PSO-RF				CSA-LSTM
	Train		Vaild		Train		Vaild		Train		Vaild
	Accuracy	F1 - score	Accuracy	F1 - score	Accuracy	F1 - score	Accuracy	F1 - score	Accuracy	F1 - score	Accuracy	F1 - score
R-CARS	0.963	0.963	0.900	0.911	0.938	0.936	0.600	0.627	0.688	0.656	0.600	0.541
R-SPA	0.625	0.604	0.500	0.330	0.663	0.664	0.200	0.157	0.650	0.634	0.500	0.363
airPLS-CARS	0.950	0.950	0.800	0.678	0.975	0.975	0.700	0.544	0.825	0.806	0.500	0.533
airPLS-SPA	0.625	0.616	0.400	0.294	0.963	0.963	0.400	0.394	0.638	0.602	0.400	0.360
SG-airPLS-CARS	0.588	0.575	0.800	0.738	0.838	0.820	0.700	0.747	0.725	0.717	0.500	0.448
SG-airPLS-SPA	0.650	0.641	0.800	0.633	0.813	0.805	0.600	0.493	0.700	0.682	0.500	0.430
SG-airPLS-(1/SG)′-CARS	0.975	0.974	0.900	0.867	0.963	0.962	0.800	0.638	0.738	0.726	0.600	0.574
SG-airPLS-(1/SG)′-SPA	0.675	0.667	0.700	0.544	0.963	0.965	0.500	0.367	0.700	0.651	0.400	0.367
SG-airPLS-MSC-CARS	0.388	0.235	0.200	0.133	0.948	0.949	0.600	0.471	0.513	0.468	0.700	0.748
SG-airPLS-MSC-SPA	0.813	0.814	0.800	0.811	0.963	0.961	0.400	0.360	0.638	0.632	0.600	0.550
SG-airPLS-SNV-CARS	0.913	0.912	0.500	0.385	0.925	0.926	0.500	0.440	0.938	0.936	0.800	0.638
SG-airPLS-SNV-SPA	0.925	0.924	0.500	0.600	0.950	0.948	0.700	0.611	0.588	0.512	0.200	0.137
SG-airPLS-SS-CARS	0.525	0.518	0.800	0.738	0.975	0.974	0.500	0.428	0.688	0.675	0.400	0.294
SG-airPLS-SS-SPA	0.800	0.775	0.700	0.643	0.963	0.963	0.600	0.653	0.750	0.705	0.400	0.380

Bold values indicate the best performance results under the corresponding algorithm.