Ensemble Classification Based on Feature Selection for Environmental Sound Recognition

Table 5

Accuracy of different feature selection methods at k = 2/3 of subsets on five classes.

Methods	Bird	Wind	Rain	Frog	Thunder
Correlation	0.6702	0.7568	0.6802	0.6184	0.702
GainRatio	0.6091	0.7461	0.6207	0.6223	0.7092
InfoGain	0.6382	0.7512	0.6321	0.6117	0.73
OneR	0.6019	0.7143	0.6419	0.6429	0.7172
ReliefF	0.6263	0.7716	0.6207	0.5924	0.6471
SymmetricalUncert	0.6443	0.7463	0.6394	0.5959	0.7206
csFs	0.6311	0.7719	0.657	0.6624	0.7534
MmEnFs1	0.74	0.775	0.765	0.62	0.75
MmEnFs2	0.755	0.785	0.76	0.665	0.7715