Predicting the growth performance of growing-finishing pigs based on net energy and digestible lysine intake using multiple regression and artificial neural networks models

Table 5 The performance of ANN models with different numbers of nodes and activation functions to predict the F/G of growing-finishing pigs¹

Number of nodes	Training data set
	Hyperbolic tangent function		Radial basis function
	R²	RMSE	R²	RMSE
1	0.797	0.37	0.816	0.35
2	0.883	0.28	0.886	0.28
3	0.905	0.25	0.898	0.26
4	0.900	0.26	0.917	0.24
5	0.918	0.23	0.928	0.22
6	0.905	0.25	0.932^*	0.21^*
7	0.917	0.24	0.910	0.25
8	0.915	0.24	0.911	0.24
9	0.900	0.26	0.907	0.25
10	0.907	0.25	0.912	0.24

RMSE root mean square error
^*Means the best performance of ANN models with different numbers of nodes and activation functions to predict F/G.
¹ All the ANN models were generated using the training data set (n = 287).

ISSN: 2049-1891