Skip to main content
. 2020 Jul 28;6(7):e04289. doi: 10.1016/j.heliyon.2020.e04289

Table 5.

Summary of the MLP neural network design approach.

Thermoplastic Material PVC
Number of datasets 42
Training dataset 36
Validation dataset 6
Training Algorithm Levenberg Marquardt
Activation Function Tansig and Purelin
Training Time 5 s
Number of Iterations 6 iterations
Performance Evaluation Mean Square Error
Number of Inputs Layers 1
Number of Input Neurons 9 (represents the properties for different grades of PVC thermoplastic)
Number of Hidden Layers 2
Number of Hidden Neurons (1st layer) 50
Number of Hidden Neurons (2nd layer) 40
Number of Output Layer 1
Number of Output Neuron 11 (represents output variable, i.e. profile settings)