Table 3 Comparison of MSE and MAE prediction performance of different methods on Milano dataset.
From: Core network traffic prediction based on vertical federated learning and split learning
| Â | Milano | |||||
|---|---|---|---|---|---|---|
Methods | MSE | MAE | ||||
| Â | SMS | Call | Internet | SMS | Call | Internet |
SVR | 0.5294 | 0.1211 | 0.1252 | 0.3981 | 0.2134 | 0.3120 |
Lasso | 0.8411 | 0.3215 | 0.4621 | 0.7214 | 0.5162 | 0.6122 |
LSTM | 0.5922 | 0.1545 | 0.1874 | 0.4721 | 0.3134 | 0.3122 |
Fedavg | 0.4853 | 0.1466 | 0.1168 | 0.4176 | 0.2045 | 0.3109 |
VFL | \(\mathbf {0.3479}\) | \(\mathbf {0.1023}\) | \(\mathbf {0.1132}\) | \(\mathbf {0.3742}\) | \(\mathbf {0.2001}\) | \(\mathbf {0.2976}\) |
