Improving Classification Performance through an Advanced Ensemble Based Heterogeneous Extreme Learning Machines

Table 2

Optimal values of ELMs’ parameters (nh: number of hidden nodes).

Dataset			nh
Iris		1.2	20
Climate		1.2	20
Credit		1.2	20
Wave		1.2	310
Satellite		1.4	310
Firm		1.2	310
Letter		1.2	700
Colon		0.3	80
Liver		1.2	20
Vowel		1.2	200