Evolution of Deep Neural Network Architecture Using Particle Swarm Optimization to Improve the Performance in Determining the Friction Angle of Soil

Table 7

Summary of the 300 simulations using R² criteria.

Model	Data set	Average	Min	Max	SD
10-HD-R^2	Training	0.946	0.915	0.970	0.0083
	Validation	0.895	0.721	0.939	0.0377
	Testing	0.935	0.861	0.966	0.0138
10-HD-MAE	Training	0.900	0.586	0.953	0.0477
	Validation	0.851	0.436	0.930	0.0657
	Testing	0.889	0.549	0.939	0.0453
8-HD-RMSE	Training	0.935	0.826	0.95	0.0156
	Validation	0.901	0.622	0.942	0.0343
	Testing	0.881	0.795	0.927	0.0186