Application of Soft Computing Techniques for the Analysis of Tractive Properties of a Low-Power Agricultural Tractor under Various Soil Conditions

Table 4

Statistics of neural model architectures used for sensitivity analysis.

Dependent variable	The range of number of neurons in hidden layer	The range of R^2 for validation dataset	The range of MAE for validation dataset	The range of RMSE for validation dataset
Traction force	16–40	0.944–0.955	0.028–0.032	0.040–0.045
Tractive efficiency	16–38	0.928–0.954	0.024–0.032	0.037–0.047