Applying Bayesian Optimization for Machine Learning Models in Predicting the Surface Roughness in Single-Point Diamond Turning Polycarbonate

Table 3

Comparison MLP-NN models under different activation functions.

Activation functions	Training dataset			Testing dataset
	RMSE	MAE		RMSE	MAE
“relu”	0.7349	0.5711	0.0765	0.6913	0.5649	0.1855
“identity”	0.3249	0.2663	0.8194	0.4829	0.4112	0.6025
“sigmoid”	0.7457	0.6447	0.0490	0.7453	0.6257	0.0533
“tanh”	0.8092	0.6965	−0.1198	0.8301	0.7007	−0.1744
“logistic”	0.7656	0.6417	−0.0023	0.7621	0.6678	0.0101