Table 5.
Diebold-Mariano test for the single-step and multi-step forecasts.
| Stock | ARIMAX-SVR | RF Regressor | LSTM | Benchmark GAN |
|---|---|---|---|---|
| Diebold-Mariano test single-step forecasts ( p -value) | ||||
| ATL | 0 | 0 | 0 | 0 |
| AZM | 0 | 0 | 0.025 | 0 |
| BZU | 0 | 0 | 0 | 0 |
| ENEL | 0 | 0 | 0 | 0 |
| ENI | 0 | 0 | 0 | 0 |
| EXO | 0 | 0 | 0 | 0 |
| G | 0 | 0 | 0 | 0 |
| IP | 0 | 0 | 0 | 0 |
| MB | 0 | 0 | 0 | 0 |
| REC | 0 | 0 | 0 | 0 |
| Diebold-Mariano test multi-step forecasts ( p -value) | ||||
| ATL | 0 | 0.2078 | 0.0064 | 0 |
| AZM | 0 | 0 | 0.0085 | 0 |
| BZU | 0 | 0 | 0.0004 | 0 |
| ENEL | 0 | 0 | 0 | 0 |
| ENI | 0 | 0 | 0 | 0 |
| EXO | 0 | 0 | 0 | 0 |
| G | 0 | 0 | 0.0562 | 0 |
| IP | 0 | 0 | 0.0012 | 0 |
| MB | 0 | 0 | 0.0081 | 0 |
| REC | 0 | 0 | 0 | 0 |
The forecasts obtained from each competitor model have been compared with those of our GAN. Text in bold denotes the non-rejection of the null hypothesis of the test (95% confidence level).