Table 3.

R² and RMSE of the dependent variables as a function of different learning algorithm and connection types.

Model	Learning Algorithms	Connections Types	Output Layer Transfer Function	Input Layer Transfer Function	Training		Testing
					fR2	gRMSE	R2	RMSE
2-6-3	BBPa	MFFFb	Hyperbolic Tangent	Hyperbolic Tangent	0.9899	0.206	0.9871	1.78
2-6-3	IBPc	MFFF	Hyperbolic Tangent	Sigmoid	0.9902	0.105	0.9900	1.49
2-7-3	IBP	MNFFd	Sigmoid	Hyperbolic Tangent	0.9906	0.06	0.9902	0651
2-7-3	QPe	MFFF	Hyperbolic Tangent	Hyperbolic Tangent	0.9710	4.025	0.9793	4.83
2-8-3	IBP	MNFFd	Sigmoid	Hyperbolic Tangent	0.9907	0.064	0.9904	0.62
2-8-3	QPe	MFFF	Sigmoid	Hyperbolic Tangent	0.9164	3.037	0.9174	3.09
2-8-3	BBP	MFFF	Hyperbolic Tangent	Sigmoid	0.9897	0.113	0.9893	1.36
2-9-3	IBP	MNFF	Sigmoid	Sigmoid	0.9915	0.0623	0.9904	0.337
2-9-3	IBP	MFFF	Sigmoid	Hyperbolic Tangent	0.9906	0.0641	0.9905	0.314

^aBatch Back Propagation;

^bMultilayer Full Feed Forward;

^cIncremental back propagation;

^dMultilayer normal Feed Forward;

^eQuick Propagation;

^fCoefficient of determination;

^gRoot mean square deviation.