A Novel Method for Classifying Body Mass Index on the Basis of Speech Signals for Future Clinical Applications: A Pilot Study

Table 6

Statistical analysis results by independent two sample -test and Benjamin-Hochberg's method.


Group	Feature	Class	Mean	Std.		value	Adj. value

Female: 20–30	aF60_120_F240_480	Normal	0.834	0.390	3.474	<0.001	0.005
		Overweight	0.699	0.365
	aF240_480_960_1960	Normal	2.285	0.818	3.510	<0.001	0.005
		Overweight	1.996	0.806
	aF60_120_960_1960	Normal	2.135	1.416	3.618	<0.001	0.005
		Overweight	1.631	1.248
	eF240_480_960_1960	Normal	3.033	0.627	3.342	<0.001	<0.01
		Overweight	2.818	0.660
	eMFCC4	Normal	1.277	6.836	2.581	<0.05	<0.05
		Overweight	−0.801	8.315
	oMFCC4	Normal	−4.087	5.624	2.757	<0.01	<0.05
		Overweight	−5.989	7.191
	SITS	Normal	56.14	7.515	3.106	<0.005	0.01
		Overweight	53.73	8.074

Male: 20–30	eMFCC4	Normal	5.057	6.678	3.393	<0.001	<0.01
		Overweight	2.679	6.929

value < 0.05 was considered statistically significant. The values were adjusted using the Benjamin-Hochberg method to control the false discovery rate. Only statistically significant features among all features selected by wrapper-based feature subset selection in each group are described in this table (Std: standard deviation, Adj: adjusted).

Evidence-Based Complementary and Alternative Medicine

A Novel Method for Classifying Body Mass Index on the Basis of Speech Signals for Future Clinical Applications: A Pilot Study

Table 6