Ensemble Classification Based on Feature Selection for Environmental Sound Recognition

Table 6

Accuracy of different feature selection methods at k = 3/4 of subsets on five classes.

Methods	Bird	Wind	Rain	Frog	Thunder
Correlation	0.6512	0.7754	0.6856	0.5829	0.6976
GainRatio	0.6215	0.7634	0.6618	0.64	0.75
InfoGain	0.6275	0.735	0.625	0.6064	0.7708
OneR	0.6436	0.7436	0.6186	0.6166	0.7333
ReliefF	0.6567	0.7143	0.648	0.6471	0.7143
SymmetricalUncert	0.63	0.7337	0.6233	0.5907	0.7668
csFs	0.6385	0.7563	0.6837	0.6493	0.7259
MmEnFs1	0.735	0.75	0.705	0.595	0.745
MmEnFs2	0.7325	0.7525	0.7175	0.655	0.75