Table 6.

The precision of drowsiness detection was obtained with the XGBoost algorithm for each subject in the training set. The shaded green cells with bold text show the highest precision obtained for each subject. At the bottom, the best mean precision 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	0.5439	0.5439	0.5439	0.5439	0.5439	0.5439	0.5439	0.5439	0.5439
tr03-0256	0.6222	0.6348	0.5676	0.5349	0.5571	0.6127	0.6096	0.4270	0.3974
tr03-0876	0.6514	0.5800	0.6765	0.5973	0.4808	0.7397	0.7123	0.5804	0.6264
tr03-1389	0.2273	0.3774	0.4906	0.5455	0.5034	0.5391	0.4556	0.5519	0.5120
tr04-0649	0.7582	0.8273	0.5733	0.7895	0.7778	0.4878	0.5294	0.9063	0.6000
tr04-0726	0.8318	1.0000	0.7128	0.6373	0.6055	0.6627	0.7333	0.8370	0.7706
tr05-1434	0.4051	0.4083	0.1429	0.6028	0.7168	0.7692	0.7805	0.6571	0.5027
tr05-1675	0.6642	0.6011	0.6885	0.5607	0.6559	0.5862	0.7265	0.5425	0.5433
tr07-0168	0.7500	0.7923	0.5759	0.5375	0.5397	0.5730	0.5963	0.4488	0.5156
tr07-0458	0.2816	0.0492	0.5437	0.4789	0.5088	0.5446	0.5732	0.4500	0.4930
tr07-0861	0.6264	0.6688	0.5338	0.5342	0.5349	0.5105	0.5521	0.6216	0.5780
tr08-0021	0.7464	0.6854	0.6119	0.4717	0.4535	0.4000	0.4688	0.6325	0.6355
tr08-0111	0.6692	0.8214	0.7297	0.8205	0.7912	0.5314	0.4840	0.6500	0.5985
tr09-0175	0.6404	0.5263	0.5742	0.4762	0.5379	0.6142	0.6053	0.5273	0.6207
tr10-0872	0.6174	0.5909	0.5045	0.5130	0.4892	0.5254	0.4882	0.5349	0.4627
tr13-0204	0.5054	0.4545	0.6357	0.4820	0.6170	0.6636	0.6348	0.6032	0.5970
Average	0.5963	0.5976	0.5691	0.5704	0.5821	0.5815	0.5934	0.5946	0.5623