Feature Selection Using Maximum Feature Tree Embedded with Mutual Information and Coefficient of Variation for Bird Sound Classification

Table 6

MICV-ERMFT compared to other methods of Kappa and DRR.

Dataset	Classifier	Method
		Cor	GR	IG	OR	RF	SU	CS	MICV	ORI
		Kappa|DRR (%)
Birds	J48	0.83 | 52	0.88 | 44	0.83 | 53	0.85 | 53	0.81 | 49	0.86 | 53	0.79 | 50	0.87 | 58	0.83 | 0
	NB	0.76 | 40	0.74 | 49	0.77 | 53	0.75 | 53	0.74 | 44	0.76 | 53	0.76 | 53	0.81 | 53	0.77 | 0
	SVM	0.92 | 52	0.92 | 44	0.89 | 53	0.87 | 44	0.92 | 53	0.88 | 53	0.92 | 53	0.93 | 57	0.92 | 0
	RFs	0.95 | 44	0.95 | 52	0.95 | 40	0.93 | 49	0.96 | 53	0.94 | 49	0.94 | 53	0.93 | 50	0.93 | 0
Crane	J48	0.64 | 44	0.74 | 50	0.63 | 50	0.70 | 53	0.63 | 40	0.62 | 50	0.63 | 53	0.76 | 53	0.70 | 0
	NB	0.73 | 50	0.75 | 52	0.71 | 44	0.73 | 53	0.72 | 50	0.71 | 50	0.69 | 46	0.76 | 53	0.75 | 0
	SVM	0.83 | 42	0.78 | 52	0.73 | 50	0.82 | 53	0.77 | 50	0.73 | 50	0.84 | 53	0.84 | 52	0.84 | 0
	RFs	0.85 | 50	0.83 | 40	0.83 | 50	0.84 | 53	0.84 | 50	0.83 | 42	0.84 | 50	0.88 | 53	0.88 | 0