A Novel Learning Scheme for Chebyshev Functional Link Neural Networks

Table 5

Comparative average performance of HCFLNN and FLNN [21] based on the confidence level (𝛼=95%).

Dataset	HCFLNN	FLNN
IRIS train	0.9964±0.0136	0.9866±0.0260
Test set	0.9864±0.0262	0.9866±0.0260
WINE train	0.9842±0.0259	0.9605±0.0405
Test set	0.9708±0.0350	0.9550±0.0431
PIMA train	0.8064±0.0395	0.7877±0.0409
Test set	0.7928±0.0405	0.7812±0.0414