Research Article
A Comparative Study of Activation Functions of NAR and NARX Neural Network for Long-Term Wind Speed Forecasting in Malaysia
Table 2
Model parameters of NAR and NARX neural networks with tansig and logsig functions.
| ā | Areas | NAR and NARX Neural Networks | Tansig | Logsig | Epoch | Time (s) | Performance | Neuron | Epoch | Time (s) | Performance | Neuron |
| NARNN | Kuala Lumpur | 160 | 58 | 1.02 | 20 | 160 | 65 | 1.55 | 20 | Kuantan | 160 | 55 | 1.21 | 20 | 160 | 65 | 1.39 | 20 | Melaka | 161 | 62 | 1.34 | 20 | 160 | 71 | 2.11 | 20 |
| NARXNN | Kuala Lumpur | 160 | 66 | 1.24 | 20 | 160 | 70 | 1.36 | .20 | Kuantan | 160 | 99 | 1.09 | 20 | 160 | 112 | 1.38 | 20 | Melaka | 160 | 78 | 1.12 | 20 | 160 | 80 | 1.29 | 20 |
|
|