An Ensemble of Adaptive Surrogate Models Based on Local Error Expectations

Table 2

Comparison results of Kriging surrogate model.

Test function	Approach	The number of initial samples	Kriging model		RBF model
			The number of total samples	Final value of R^2	The number of total samples	Final value of R^2
BN	LEE	10	18	0.946	15	0.908
	MSDA		27	0.899	27	0.873
H3	LEE	15	26	0.896	35	0.902
	MSDA		34	0.837	39	0.879
CV	LEE	20	29	0.909	30	0.934
	MSDA		44	0.943	41	0.901
SHC	LEE	10	25	0.920	21	0.941
	MSDA		36	0.883	29	0.866