Table 4. Analysing machine learning... | Emerald Publishing

Table 4.

Analysing machine learning models’ performance for CBR value prediction on training and testing sets

Model	Symbol	Data set	R-squared	MA_Error	RMS_Error	RA_Error	RRS_Error
Random forest	RF	Training set	0.9999	0.1271	0.1617	1.2316%	1.3592%
	—	Testing set	0.9997	0.1992	0.2782	2.0942%	2.4832%
Artificial neural network	ANN	Training set	0.9989	0.4052	0.5588	3.9260%	4.6975%
	—	Testing set	0.9986	0.5147	0.6589	5.4104%	5.8807%
Support vector machine	SVM	Training set	0.9972	0.6924	0.8982	6.7079%	7.5512%
	—	Testing set	0.9968	0.7060	0.9380	7.4214%	8.3714%
Gaussian process regression	GPR	Training set	0.9941	1.031	1.3390	9.9880%	11.2571%
	—	Testing set	0.9931	1.0894	1.5060	11.4518%	13.4408%

Model	Symbol	Data set	R-squared	MA_Error	RMS_Error	RA_Error	RRS_Error
Random forest	RF	Training set	0.9999	0.1271	0.1617	1.2316%	1.3592%
	—	Testing set	0.9997	0.1992	0.2782	2.0942%	2.4832%
Artificial neural network	ANN	Training set	0.9989	0.4052	0.5588	3.9260%	4.6975%
	—	Testing set	0.9986	0.5147	0.6589	5.4104%	5.8807%
Support vector machine	SVM	Training set	0.9972	0.6924	0.8982	6.7079%	7.5512%
	—	Testing set	0.9968	0.7060	0.9380	7.4214%	8.3714%
Gaussian process regression	GPR	Training set	0.9941	1.031	1.3390	9.9880%	11.2571%
	—	Testing set	0.9931	1.0894	1.5060	11.4518%	13.4408%