. 2023 Mar 1;20(5):4430. doi: 10.3390/ijerph20054430

Table 3.

Parameters with the highest accuracy (mean ± standard deviation), when comparing the groups of patients SG vs FG.

Data Base	Classification Techniques	Reduction Dimensionality Techniques	Parameters Selected	Accuracy (%)
Dataset1	NN	Forward Selection	PF(f/V_T), P(V_T/T_I), IQR(T_I/T_Tot), PA(f/V_T)	65.30 ± 5.9
Dataset1	NN	Bidirectional Search	PF(f/V_T), P(V_T/T_I), IQR(T_I/T_Tot)	54.40 ± 6.9
Dataset2	NN	Forward Selection	IQR(CD1-T_I), K(CA8-RR)	84.61 ± 3.1
Dataset2	NN	Bidirectional Search	IQR(CD1-T_I), K(CA8-RR), IQR(CD1-V_T/T_I), K(CD1-T_I), $\bar{X}$ (CA7-f/V_T)	80.76 ± 6.2
Dataset1	LDA	Forward Selection	PA(f/V_T)	70.00 ± 7.8
Dataset1	LDA	Bidirectional Search	PA(f/V_T), P(f/V_T)	70.60 ± 7.7
Dataset2	LDA	Forward Selection	IQR(CD2-T_I/T_Tot)	71.64 ± 0.8
Dataset2	LDA	Bidirectional Search	IQR(CD2-T_I/T_Tot) $\bar{X}$ (CA7-f/V_T), IQR(CD1-V_T/T_I)	72.13 ± 4.8