New Zealand Journal of Forestry Science

Table 8 Performance metrics of linear models estimating MOE of logs. Italicised values indicate the recommended model; bold values indicate the best value(s) in each column

From: Validation of models predicting modulus of elasticity in Douglas-fir trees, boles, and logs

	Calibration dataset						Validation dataset			Comments
Fixed effects^a	R ² _marg	R ² _cond	AIC	RMSE	MAPE	TP rate	RMSE	MAPE	TP rate
				(GPa)	(%)	(%)	(GPa)	(%)	(%)
V ²	0.43	0.63	1513	1.3	10	61	1.4	11	74	Best 1-variable
D	0.02	0.46	1695	1.9	15	2	1.7	14	0	Worst
H	0.06	0.58	1647	1.8	14	0	1.7	13	0
A	0.09	0.55	1649	1.7	14	21	1.6	13	17
E	0.51	0.57	1501	1.2	10	64	1.4	10	62
V ² + ρ	0.53	0.61	1485	1.2	9	67	*1.4*	10	74	Best validation
V ² + T	0.44	0.64	1498	1.3	10	62	1.4	11	71
V ² + A	0.43	0.63	1509	1.3	10	61	1.4	11	71
E + D	0.53	0.65	1471	1.2	9	64	1.4	10	60
E + T	0.54	0.59	1480	1.2	9	64	1.4	11	57
E + A	0.62	0.69	1424	1.1	9	71	1.4	10	69	Best equal AIC
E + H	0.56	0.69	1436	1.2	9	64	1.4	10	67
A + T	0.12	0.53	1636	1.7	14	33	1.5	12	36
A + ρ	0.12	0.56	1645	1.7	13	27	1.6	13	21
H + ρ	0.08	0.62	1638	1.8	14	22	1.7	13	12
D + A	0.25	0.54	1621	1.6	13	39	1.6	13	45
D + H	0.13	0.60	1626	1.7	14	19	1.6	13	21
D + ρ + H	0.17	0.62	1610	1.7	13	27	1.7	13	26
D + ρ + A	0.31	0.54	1611	1.5	12	45	1.6	13	50
V ² + ρ + D	0.54	0.67	1469	1.2	9	67	1.4	10	69
V ² + ρ + T	0.56	0.67	1459	1.2	9	68	1.4	11	74	Similar to V ²
V ² + ρ + A	0.63	0.70	1425	1.1	9	73	1.4	11	69
V ² + ρ + H	0.55	0.69	1439	1.2	9	65	1.4	11	69
V ² + ρ + A + D	0.64	0.70	1424	1.1	9	70	1.5	11	69	Best equal AIC

^aAll variables are as defined in Table 2

Back to article page