Research Article
Particle Swarm Optimization-Based Support Vector Regression for Tourist Arrivals Forecasting
Table 5
Training results of GRIDSVR, PSOSVR, and FS–PSOSVR.
| Model | Data set | C | ε | σ |
| GRIDSVR | Japan | 1024.00 | 0.0313 | 0.0156 | Hong Kong and Macao | 128.00 | 0.0039 | 0.0313 | South Korea | 1024.00 | 0.0156 | 0.0039 | The United States | 128.00 | 0.2500 | 0.0625 | Total | 256.00 | 0.0039 | 0.0039 |
| PSOSVR | Japan | 805.03 | 0.0423 | 0.0130 | Hong Kong and Macao | 209.96 | 0.0039 | 0.0367 | South Korea | 776.87 | 0.0159 | 0.0039 | The United States | 415.75 | 0.0467 | 0.0367 | Total | 216.66 | 0.0039 | 0.0039 |
| FS–PSOSVR | Japan | 564.45 | 0.1371 | 0.0241 | Hong Kong and Macao | 389.08 | 0.0039 | 0.0039 | South Korea | 673.88 | 0.0222 | 0.0094 | The United States | 531.61 | 0.0268 | 0.0405 | Total | 1016.71 | 0.0087 | 0.0060 |
|
|