Research Article
Vessel Track Prediction Based on Fractional Gradient Recurrent Neural Network with Maneuvering Behavior Identification
Table 8
Testing and training accuracy with different fractional orders and batch sizes.
| ā | 200 | 300 | 400 |
| | Testing | Training | Testing | Training | Testing | Training | 0.9 | 94.350000% | 95.110000% | 95.590000% | 96.000000% | 95.780000% | 96.478333% | 0.8 | 95.100000% | 95.791667% | 95.500000% | 96.326667% | 95.210000% | 96.103333% | 0.7 | 95.130000% | 95.683333% | 95.800000% | 96.305000% | 95.750000% | 96.450000% | 0.6 | 95.050000% | 95.708333% | 95.660000% | 96.345000% | 95.430000% | 96.188333% | 0.5 | 93.550000% | 94.170000% | 95.800000% | 96.385000% | 95.130000% | 96.215000% | 0.4 | 94.730000% | 95.641667% | 95.610000% | 96.343333% | 95.670000% | 96.533333% | 0.3 | 95.550000% | 96.373333% | 95.810000% | 96.668333% | 95.820000% | 96.696667% | 0.2 | 95.280000% | 96.091667% | 95.960000% | 96.720000% | 96.030000% | 96.866667% | 0.1 | 95.500000% | 96.278333% | 95.580000% | 96.416667% | 96.050000% | 96.876667% |
|
|