Table 5.

Results of the training and test phases of the Gaussian process regression using different kernel implementations. Results are presented as root-mean-square error (RMSE), mean absolute error (MAE), Pearson correlation coefficient $\rho$ and the balanced accuracy (BA).

Phase	Kernel	RMSE [1]	MAE [1]	Pearson’s $\mathit{\rho }$ [1]	BA [%]
Training/ Validation	quadratic	0.45952	0.34478	0.62269	96.37
	Matern 5/2	0.46884	0.34977	0.60161	95.12
	exponential	0.39621	0.29826	0.75242	98.01
Test	quadratic	0.46909	0.35435	0.60126	94.74
	Matern 5/2	0.47092	0.35226	0.59710	93.73
	exponential	0.46800	0.35582	0.60365	94.99