Innovative deep learning artificial intelligence applications for predicting relationships between individual tree height and diameter at breast height

Ercanlı, İlker

doi:10.1186/s40663-020-00226-3

Forest Ecosystems

Table 3 The goodness-of-fit statistics r, AAE, max. AE, RMSE, RMSE%, Bias, Bias%, FI, AIC and BIC for the best predictive DLA models with best predictive number of neuron alternative according to each hidden layer choices, the ITH-DBH functions based on NLRM, M5 based on NLME with f random, FFB-ANN and CC-ANN

From: Innovative deep learning artificial intelligence applications for predicting relationships between individual tree height and diameter at breast height

AI alternatives and regression models	The number of hidden layers	RMSE	RMSE% (%)	AIC	BIC	FI	AAE	Max. AE	Bias	Bias% (%)
50 # neuron	3	0.6416	5.6974	− 757.2286	1126.3846	0.9253	0.5051	2.3653	0.0701	0.6225
90 # neuron	4	0.6110	5.4257	− 841.2650	1042.3481	0.9323	0.4620	2.3596	0.0391	0.3469
90 # neuron	5	0.5942	5.2766	− 889.1954	994.4178	0.9359	0.4474	2.9231	−0.0111	−0.0988
100 # neuron	6	0.5859	5.2021	−913.6424	969.9708	0.9377	0.4346	2.5950	0.0447	0.3973
70 # neuron	7	0.5940	5.2747	−889.8224	993.7907	0.9360	0.4531	2.4365	0.0053	0.0468
90 # neuron	8	0.5798	5.1482	− 931.5577	952.0554	0.9390	0.4336	2.3822	0.0263	0.2339
100 # neuron	9	0.5575	4.9504	−998.9540	884.6591	0.9436	0.4077	2.5106	0.0057	0.0502
80 # neuron	10	0.5892	5.2315	−903.9603	979.6528	0.9370	0.4286	2.5746	0.0512	0.4546
M1 NLRM		0.8193	7.2746	−336.8843	1546.7288	0.8782	0.6058	4.4859	−0.2465	−2.1891
M2 NLRM		0.8224	7.3027	− 330.2519	1553.3612	0.8773	0.6102	4.3628	−0.2609	−2.3169
M3 NLRM		0.8306	7.3750	−313.3060	1570.3072	0.8749	0.6170	4.2243	−0.2695	−2.3927
M4 NLRM		0.7647	6.7905	−455.3473	1428.2659	0.8939	0.6064	4.0474	−0.0020	− 0.0174
M5 NLRM		0.7621	6.7672	−461.2447	1422.3685	0.8946	0.6132	3.8927	−0.0005	−0.0047
M6 NLRM		0.7922	7.0343	−394.6698	1488.9433	0.8862	0.6254	4.1908	−0.0022	− 0.0192
M7 NLRM		0.8158	7.2440	− 344.1317	1539.4815	0.8793	0.6070	4.4574	−0.2540	−2.2551
NLME for M5 with f random		0.7073	6.2807	−589.5670	1294.0461	0.9092	0.5769	3.4091	−0.0006	−0.0050
FFB-ANN based on A3 activation function alternative and 85 # neuron		0.7160	6.3576	−568.6345	1314.9787	0.9070	0.5775	2.5863	−0.0175	−0.1557
CC-ANN based on A3 activation function alternative and 73 # neuron		0.7110	6.3132	−580.6896	1302.9236	0.9083	0.5638	2.9023	−0.0144	−0.1282

Back to article page