Research Article
Sensitivity Analysis of the Artificial Neural Network Outputs in Friction Stir Lap Joining of Aluminum to Brass
Table 2
Summary of different networks evaluated to yield the criteria of network performance.
| No. | Activation function |
Neurons in hidden layer |
Training MRE |
Testing MRE | Hidden layer | Output layer |
| 1 | logsig | logsig | 4 | 1.6 | 2.8 | 2 | logsig | logsig | 5 | 1.2 | 3.7 | 3 | logsig | logsig | 6 | 0.45 | 2.1 | 4 | logsig | purline | 5 | 2 | 4.5 | 5 | logsig | purline | 6 | 1.63 | 3.52 | 6 | logsig | purline | 7 | 1.27 | 3.81 | 7 | logsig | tansig | 6 | 1.54 | 2.87 | 8 | logsig | tansig | 7 | 2.67 | 2.34 | 9 | purlin | logsig | 6 | 0.89 | 2.9 | 10 | purlin | logsig | 7 | 0.75 | 2.3 | 11 | tansig | purline | 6 | 1.35 | 2.74 | 12 | tansig | purline | 7 | 1.72 | 3.37 |
|
|