Table 4.

The classification accuracy was obtained with the XGBoost algorithm for each subject in the training set. The shaded green cells with bold text show the highest accuracy obtained for each subject. At the bottom, the best mean accuracy for each ratio index is marked in the same way.

Subject	Index1	Index2	θ/α	β/α	(θ + α)/β	θ/β	(θ + α)/(α + β)	γ/δ	(γ + β)/(δ + α)
	This Work		Eoh et al. [36]		Jap et al. [37]			da Silveira et al. [10]
tr03-0092	0.5420	0.5252	0.6387	0.5168	0.5504	0.6092	0.6050	0.4684	0.5527
tr03-0256	0.5924	0.6303	0.5840	0.5504	0.5672	0.6345	0.6345	0.4473	0.4346
tr03-0876	0.6387	0.5672	0.6008	0.6218	0.4748	0.6471	0.6303	0.5781	0.5992
tr03-1389	0.3487	0.3908	0.4874	0.5588	0.5042	0.5378	0.4664	0.5696	0.5148
tr04-0649	0.6975	0.8025	0.5462	0.5462	0.5630	0.4832	0.5210	0.6118	0.5527
tr04-0726	0.7983	0.7605	0.6681	0.6176	0.5966	0.6134	0.6765	0.7637	0.7511
tr05-1434	0.3739	0.3697	0.4160	0.6218	0.7059	0.5882	0.6933	0.5485	0.5063
tr05-1675	0.6849	0.6513	0.6933	0.5546	0.6218	0.5630	0.7227	0.5781	0.5781
tr07-0168	0.7773	0.8193	0.6008	0.5504	0.5630	0.6092	0.6303	0.4473	0.5105
tr07-0458	0.3109	0.2689	0.5378	0.4748	0.5084	0.5378	0.5504	0.4684	0.4979
tr07-0861	0.6933	0.7269	0.5378	0.5420	0.5504	0.5168	0.5714	0.6540	0.6160
tr08-0021	0.7857	0.6387	0.5630	0.4874	0.4664	0.3866	0.4748	0.6329	0.6245
tr08-0111	0.6891	0.8025	0.7143	0.7101	0.7227	0.5462	0.4748	0.6287	0.6118
tr09-0175	0.6050	0.5252	0.5966	0.4748	0.5420	0.6218	0.6008	0.5401	0.6498
tr10-0872	0.6134	0.6008	0.5084	0.5168	0.4874	0.5252	0.4832	0.5274	0.4810
tr13-0204	0.5042	0.4538	0.6597	0.4790	0.6387	0.6471	0.6303	0.5570	0.5570
Average	0.6035	0.5959	0.5846	0.5515	0.5664	0.5667	0.5853	0.5638	0.5649