Developing a Robust Model Based on the Gaussian Process Regression Approach to Predict Biodiesel Properties

Table 3

The comparison between four kernel functions for CP property from the point of view of statistical quantities.

Model	Phase	R^2	MRE (%)	MSE	RMSE	STD
GPR (Exponential)	Train	0.967	0.00069	0.00001	0.00247	0.00158
	Test	0.966	0.37589	1.92262	1.38659	0.93954
	Total	0.955	0.09442	0.48066	1.38659	0.64771
GPR (Matern)	Train	0.999	0.07276	0.06294	0.25089	0.15196
	Test	0.874	0.62866	7.30808	2.70335	2.15533
	Total	0.944	0.37087	2.73100	2.70335	1.30838
GPR (Squared Exponential)	Train	0.999	0.04991	0.03376	0.18375	0.12258
	Test	0.904	0.79289	9.58471	3.09592	2.24565
	Total	0.950	0.27415	2.50823	3.09592	1.40005
GPR (Rational Quadratic)	Train	1.000	0.31099	1.38429	1.17656	0.81684
	Test	0.950	0.69381	5.80117	2.40856	1.48864
	Total	0.990	0.39714	2.40895	2.40856	1.10468