Application of Gene Expression Programming to Evaluate Strength Characteristics of Hydrated-Lime-Activated Rice Husk Ash-Treated Expansive Soil

Table 5

Statistical calculations, performance indices, and objective functions of the GEP models for CBR, UCS, and R value.

GEP model	Dataset	Statistical parameters
		MAE	RSE	RMSE	NSE	R	RRMSE	ρ	OBF
CBR	Training	0.5	0.003	4.94	0.997	0.998	0.202	0.101	0.028
	Testing	0.3	0.011	3.69	0.989	0.996	0.271	0.136
	Validation	0.5	0.011	5.49	0.989	0.996	0.167	0.084
UCS28	Training	7.8	0.151	13.06	0.849	0.924	0.076	0.040	0.013
	Testing	2.7	0.970	12.21	0.903	0.971	0.081	0.041
	Validation	13.7	0.647	13.61	0.353	0.624	0.067	0.041
RValue	Training	0.2	0.003	4.49	0.997	0.998	0.222	0.111	0.032
	Testing	0.2	0.007	4.22	0.992	0.997	0.238	0.119
	Validation	0.1	0.012	4.72	0.988	0.994	0.193	0.097