Analysis of plot-level volume increment models developed from machine learning methods applied to an uneven-aged mixed forest

Table 2 Characteristics of RBF and MLP-based ANNs and associated metrics for model training

Index	Structure	Algorithm	Error function	Hidden activation	R²	RMSE	%RMSE	BIAS	%BIAS
1	MLP 7-12-1	BFGS 11	SOS	Logistic	0.908	0.5029	20.867	0.0011	0.0047
2	MLP 7-5-1	BFGS 14	SOS	Identity	0.908	0.5020	20.832	0.0468	0.1971
3	MLP 7-9-1	BFGS 11	SOS	Identity	0.908	0.5020	20.832	0.0011	0.0047
4	MLP 7-10-1	BFGS 16	SOS	Identity	0.907	0.5041	20.917	0.0468	0.1971
5	MLP 7-11-1	BFGS 11	SOS	Tanh	0.911	0.4945	20.52	0.0009	0.0016
6	RBF 7-24-1	RBFT	SOS	Gaussian	0.854	0.624	25.893	0.0010	0.0020
7	RBF 7-30-1	RBFT	SOS	Gaussian	0.895	0.534	22.179	0.0450	0.0863
8	RBF 7-27-1	RBFT	SOS	Gaussian	0.865	0.601	24.950	0.0010	0.0020
9	RBF 7-30-1	RBFT	SOS	Gaussian	0.877	0.576	23.905	0.0450	0.0863
10	RBF 7-29-1	RBFT	SOS	Gaussian	0.810	0.704	29.213	0.0010	0.0020

ISSN: 1297-966X