표 4. | Table 4. 실험 결과(F1-score) | Experimental result (F1-score)

특징 조합	FS	#F	SVM	MLP
MFCCs (baseline)	-	13	53.69	46.52
MFCCs+음질	-	13+5	52.73	51.10
	RFE	13+2	53.31	48.29
	ETC	13+2	53.31	48.29
MFCCs+운율	-	13+20	59.80	56.14
	RFE	13+8	58.8	51.11
	ETC	13+7	57.51	50.18
MFCCs+발음	-	13+7	74.45	64.45
	RFE	13+3	73.23	66.56
	ETC	13+5	72.03	67.37
MFCCs+음질+운율	-	13+25	60.09	59.96
	RFE	13+11	58.97	48.19
	ETC	13+9	63.42	54.94
MFCCs+음질+발음	-	13+12	72.89	68.79
	RFE	13+6	76.87	73.85
	ETC	13+6	76.77	73.02
MFCCs+운율+발음	-	13+27	74.20	66.17
	RFE	13+12	76.47	74.91
	ETC	13+9	78.13	68.22
MFCCs+음질+운율 +발음	-	13+32	69.41	69.13
	RFE	13+14	77.28	65.51
	ETC	13+10	80.15	78.09

SVM, support vector machine; MLP, multiple layer perceptron; MFCCs, mel frequency cepstral coefficients; RFE, recursive feature elimination; ETC, extra trees classifier.