Evaluating soil nutrients of Dacrydium pectinatum in China using machine learning techniques

Forest Ecosystems

Table 4 Artificial neural network calculation results

Algorithm	Training set	Testing set	RMSE	MSPE	ME
BPNN	210	45	5.6	31	−0.59
FPNN	210	45	6.8	42	−0.53
MLPNN	210	45	7.2	55	−0.51
GRNN	210	45	5.1	29	−0.85

Note: BPNN is the back-propagation neural network; FPNN is the field probing neural network; MLPNN is the multilayer-propagation feed forward neural network; GRNN is the generalized regression neural network; MSPE is the mean square prediction error; RMSE is the root-mean-squared error; ME is the mean error