Optimum model parameters and results
| Classification prediction model | ||
|---|---|---|
| Model | Reference DBN | Tent-SSA-DBN |
| Number of nodes in the 1st hidden layer | 200 | 261 |
| Number of nodes in the 2nd hidden layer | 50 | 61 |
| Learning rate | 0.01 | 0.009174 |
| Random seed | 7 | 0 |
| Accuracy | 94.23% | 99.99% |
| Classification prediction model | ||
|---|---|---|
| Model | Reference DBN | Tent-SSA-DBN |
| Number of nodes in the 1st hidden layer | 200 | 261 |
| Number of nodes in the 2nd hidden layer | 50 | 61 |
| Learning rate | 0.01 | 0.009174 |
| Random seed | 7 | 0 |
| Accuracy | 94.23% | 99.99% |
| Regression forecasting model | |||
|---|---|---|---|
| Model | Reference DBN | Tent-SSA-DBN (Based on RMSE) | Tent-SSA-DBN (Based on MaxAE) |
| Number of nodes in the 1st hidden layer | 200 | 78 | 328 |
| Number of nodes in the 2nd hidden layer | 50 | 75 | 47 |
| Learning rate | 0.01 | 0.025446 | 0.009279 |
| Random seed | 7 | 443 | −1,116 |
| RMSE | 0.2161 | 0.1264 | 0.1459 |
| MAE | 0.1701 | 0.0934 | 0.1228 |
| Rˆ2 | 0.9070 | 0.9682 | 0.9576 |
| MaxAE | 0.5157 | 0.4149 | 0.3013 |
| Optimization effect | - | 41.51% | 41.57% |
| Regression forecasting model | |||
|---|---|---|---|
| Model | Reference DBN | Tent-SSA-DBN (Based on RMSE) | Tent-SSA-DBN (Based on MaxAE) |
| Number of nodes in the 1st hidden layer | 200 | 78 | 328 |
| Number of nodes in the 2nd hidden layer | 50 | 75 | 47 |
| Learning rate | 0.01 | 0.025446 | 0.009279 |
| Random seed | 7 | 443 | −1,116 |
| RMSE | 0.2161 | 0.1264 | 0.1459 |
| MAE | 0.1701 | 0.0934 | 0.1228 |
| Rˆ2 | 0.9070 | 0.9682 | 0.9576 |
| MaxAE | 0.5157 | 0.4149 | 0.3013 |
| Optimization effect | - | 41.51% | 41.57% |
Source(s): Authors' own work