Study of a Vocal Feature Selection Method and Vocal Properties for Discriminating Four Constitution Types

Table 1

92 features selected as having less than 20% CV in each of the six subjects.

Features
aF0	oJITT	sF10	uF1_ln
aF2/aF1	oF2/oF1	sF50	aF2_ln
aF3/aF1	oF3/oF1	sF90	eF2_ln
aF4/aF1	oF4/oF1	sFHL	iF2_ln
aF3/aF2	oF3/oF2	sI10	oF2_ln
aF4/aF2	oF4/oF2	sI50	uF2_ln
aF4/aF3	oF4/oF3	sI90	aF3_ln
eF0	uT0	sIHL	eF3_ln
eF2/eF1	uPPQ	sF0	iF3_ln
eF3/eF1	uF2/uF1	sFSTD	oF3_ln
eF4/eF1	uF3/uF1	sI0	uF3_ln
eF3/eF2	uF4/uF1	sISTD	aF4_ln
eF4/eF2	uF3/uF2	iFB240_480/iFB960_1920	eF4_ln
eF4/eF3	uF4/uF2	oFB240_480/oFB960_1920	iF4_ln
iF0	uF4/uF3	uFB240_480/uFB960_1920	oF4_ln
iT0	aMFCC13	sFB60_120/sFB240_480	uF4_ln
iF2/iF1	eMFCC2	sFB240_480/sFB960_1920	eT0_ln
iF3/iF1	eMFCC13	sFB60_120/sFB960_1920	sSPD_ln
iF4/iF1	iMFCC2	uF0_ln	aJITA_sqrt
iF3/iF2	iMFCC13	aF1_ln	eJITA_sqrt
iF4/iF3	oMFCC6	eF1_ln	iJITA_sqrt
oF0	oMFCC13	iF1_ln	oJITA_sqrt
oT0	uMFCC13	oF1_ln	uJITA_sqrt