Table 4.
Multi-step (5 days) forecast errors for ARIMAX, RF Regressor, LSTM, the benchmark GAN, and our GAN architecture.
| Multi-step forecasts | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Stock | Metrics | ARIMAX-SVR | RF Regressor | LSTM | Benchmark GAN | Our GAN | ||||
| (p,d,q) | Results | Results | Lags | Results | Lags | Results | Lags | Results | ||
| ATL | RMSE | (5,1,0) | 1.784 | 1.027 (0.015) | 5 | 1.080 (0.021) | 2 | 2.319 (0.543) | 2 | 0.950 (0.027) |
| MAE | 1.377 | 0.694 (0.009) | 0.702 (0.014) | 1.782 (0.424) | 0.620 (0.021) | |||||
| MAPE | 8.267 | 3.833 (0.052) | 3.920 (0.074) | 9.409 (2.393) | 3.287 (0.118) | |||||
| AZM | RMSE | (5,1,0) | 1.776 | 0.977 (0.027) | 5 | 0.930 (0.018) | 2 | 2.274 (0.635) | 2 | 0.850 (0.024) |
| MAE | 1.380 | 0.697 (0.013) | 0.647 (0.013) | 2.050 (0.540) | 0.588 (0.020) | |||||
| MAPE | 9.394 | 4.444 (0.088) | 4.172 (0.083) | 14.026 (2.809) | 3.844 (0.176) | |||||
| BZU | RMSE | (4,1,0) | 1.799 | 0.906 (0.016) | 5 | 0.886 (0.026) | 2 | 2.444 (0.651) | 2 | 0.784 (0.030) |
| MAE | 1.549 | 0.669 (0.007) | 0.654 (0.018) | 1.984 (0.603) | 0.572 (0.022) | |||||
| MAPE | 8.117 | 3.489 (0.039) | 3.443 (0.095) | 11.179 (4.202) | 2.951 (0.108) | |||||
| ENEL | RMSE | (2,1,0) | 0.607 | 0.563 (0.005) | 5 | 0.388 (0.016) | 2 | 1.647 (0.622) | 2 | 0.268 (0.023) |
| MAE | 0.449 | 0.357 (0.005) | 0.267 (0.010) | 1.283 (0.472) | 0.182 (0.020) | |||||
| MAPE | 6.379 | 4.829 (0.096) | 3.871 (0.113) | 20.652 (4.301) | 2.771 (0.279) | |||||
| ENI | RMSE | (1,1,0) | 1.649 | 2.114 (0.013) | 5 | 1.325 (0.081) | 2 | 2.041 (0.624) | 2 | 0.888 (0.193) |
| MAE | 1.200 | 1.366 (0.012) | 0.909 (0.053) | 1.418 (0.437) | 0.636 (0.134) | |||||
| MAPE | 12.514 | 16.028 (0.121) | 10.288 (0.639) | 12.156 (3.720) | 5.884 (1.164) | |||||
| EXO | RMSE | (1,1,0) | 5.201 | 4.840 (0.013) | 5 | 4.116 (0.071) | 2 | 5.721 (1.044) | 2 | 2.955 (0.218) |
| MAE | 3.691 | 3.363 (0.013) | 3.047 (0.030) | 4.776 (0.870) | 2.264 (0.258) | |||||
| MAPE | 6.664 | 5.461 (0.022) | 5.195 (0.034) | 9.061 (1.909) | 3.977 (0.461) | |||||
| G | RMSE | (5,1,0) | 0.526 | 0.622 (0.018) | 5 | 0.564 (0.004) | 2 | 0.860 (0.338) | 2 | 0.489 (0.003) |
| MAE | 0.369 | 0.441 (0.010) | 0.385 (0.006) | 0.721 (0.288) | 0.339 (0.010) | |||||
| MAPE | 2.445 | 3.000 (0.007) | 2.641 (0.035) | 4.763 (1.879) | 2.281 (0.076) | |||||
| IP | RMSE | (1,1,0) | 2.057 | 4.363 (0.017) | 5 | 2.557 (0.089) | 2 | 2.866 (0.716) | 2 | 1.366 (0.155) |
| MAE | 1.674 | 2.350 (0.013) | 1.670 (0.033) | 2.404 (0.603) | 1.007 (0.139) | |||||
| MAPE | 5.761 | 6.629 (0.028) | 5.178 (0.080) | 7.941 (1.989) | 3.577 (0.527) | |||||
| MB | RMSE | (5,1,0) | 0.680 | 0.474 (0.020) | 5 | 0.419 (0.002) | 2 | 0.880 (0.566) | 2 | 0.361 (0.012) |
| MAE | 0.508 | 0.335 (0.013) | 0.289 (0.004) | 0.665 (0.354) | 0.257 (0.011) | |||||
| MAPE | 6.738 | 4.416 (0.183) | 3.836 (0.055) | 7.059 (2.086) | 3.300 (0.143) | |||||
| REC | RMSE | (2,1,0) | 3.742 | 3.231 (0.096) | 5 | 2.356 (0.117) | 2 | 4.736 (1.196) | 2 | 1.776 (0.261) |
| MAE | 3.278 | 2.191 (0.065) | 1.625 (0.286) | 3.875 (1.047) | 1.403 (0.236) | |||||
| MAPE | 9.320 | 5.295 (0.155) | 4.427 (0.174) | 10.886 (2.780) | 3.713 (0.626) | |||||
Text in bold denotes the best results (95% confidence level).