Predicting and Investigating the Permeability Coefficient of Soil with Aided Single Machine Learning Algorithm

Table 3

Min, max, average, and StD of performance values of 10000 performance values.

		R^2				RMSE (×10^−11 m/s)
		StD	Min	Average	Max	StD	Min	Average	Max
Training	SVM	0.046	0.541	0.745	0.884	0.090	0.428	0.754	0.943
	KNN	0.047	0.613	0.765	0.935	0.107	0.252	0.711	0.927
	LightGBM	0.057	0.380	0.587	0.784	0.097	0.530	0.938	1.192
	RF	0.009	0.884	0.947	0.972	0.031	0.237	0.342	0.438
	GB	0.000	0.998	0.999	1.000	0.007	0.019	0.036	0.067
Testing	SVM	0.092	0.597	0.740	0.960	0.234	0.148	0.740	1.333
	KNN	0.085	0.549	0.715	0.961	0.208	0.173	0.801	1.403
	LightGBM	0.088	0.342	0.522	0.819	0.196	0.460	1.054	1.749
	RF	0.055	0.721	0.801	0.960	0.167	0.209	0.658	1.036
	GB	0.054	0.724	0.804	0.971	0.153	0.199	0.641	1.013
		MAE (×10^−11 m/s)				MAPE (%)
Training	SVM	0.042	0.214	0.351	0.457	0.028	0.156	0.252	0.356
	KNN	0.058	0.180	0.381	0.526	0.028	0.156	0.252	0.356
	LightGBM	0.083	0.346	0.666	0.931	0.112	0.411	0.793	1.191
	RF	0.025	0.087	0.177	0.243	0.017	0.087	0.138	0.196
	GB	0.005	0.014	0.028	0.051	0.006	0.017	0.035	0.061
Testing	SVM	0.109	0.118	0.394	0.761	0.045	0.147	0.258	0.703
	KNN	0.108	0.137	0.446	0.876	0.055	0.155	0.293	0.602
	LightGBM	0.134	0.309	0.756	1.267	0.247	0.297	0.855	1.939
	RF	0.084	0.163	0.396	0.680	0.074	0.167	0.324	0.784
	GB	0.081	0.161	0.390	0.656	0.077	0.157	0.321	0.830