Improved Instance Selection Methods for Support Vector Machine Speed Optimization

Table 8

Statistical analysis, classification accuracy.

Technique	Number of samples	Standard variation	-test values ()
EISA versus PSC	10	7.51404	−2.8953
EISA versus DROP 3	10	7.51404	−0.1641
EISA versus DROP 5	10	7.51404	−0.0462
EISA versus GCNN	10	7.51404	−2.0158
EISA versus POCNN	10	7.51404	−1.2457
FFA versus PSC	10	3.5072	−10.4952
FFA versus DROP 3	10	3.5072	−4.6438
FFA versus DROP 5	10	3.5072	−4.3911
FFA versus GCNN	10	3.5072	−9.0616
FFA versus POCNN	10	3.5072	−7.4116
FFA_Clus versus PSC	10	0.7447	−86.626
FFA_Clus versus DROP 3	10	0.7447	−59.0670
FFA_Clus versus DROP 5	10	0.7447	−57.8781
FFA_Clus versus GCNN	10	0.7447	−79.8520
FFA_Clus versus POCNN	10	0.7447	−72.1035