Research Article
Evolutionary Spiking Neural Networks for Solving Supervised Classification Problems
Table 3
Accuracy of design and testing on every configuration for Balance Scale, Blood, Breast Cancer, Card, Diabetes, and Fertility datasets.
| Dataset | Configuration | Design accuracy | Test accuracy |
| Balance Scale | α1 | 0.7486 ± 0.0525 | 0.7219 ± 0.0629 | α2 | 0.8354 ± 0.0460 | 0.8077 ± 0.0582 | β1 | 0.7331 ± 0.0718 | 0.6944 ± 0.0752 | β2 | 0.8363 ± 0.0272 | 0.8078 ± 0.0427 | γ1 | 0.8528 ± 0.0197 | 0.8346 ± 0.0261 | γ2 | 0.8960 ± 0.0062 | 0.8647 ± 0.0134 |
| Blood | α1 | 0.7711 ± 0.0112 | 0.7622 ± 0.0117 | α2 | 0.8010 ± 0.0135 | 0.7731 ± 0.0168 | β1 | 0.7747 ± 0.0110 | 0.7607 ± 0.0138 | β2 | 0.7863 ± 0.0162 | 0.7684 ± 0.0120 | γ1 | 0.7760 ± 0.0076 | 0.7618 ± 0.0088 | γ2 | 0.7957 ± 0.0145 | 0.7685 ± 0.0155 |
| Breast Cancer | α1 | 0.9494 ± 0.0141 | 0.9418 ± 0.0238 | α2 | 0.9781 ± 0.0073 | 0.9585 ± 0.0077 | β1 | 0.9474 ± 0.0121 | 0.9405 ± 0.0151 | β2 | 0.9677 ± 0.0073 | 0.9432 ± 0.0142 | γ1 | 0.9574 ± 0.0111 | 0.9384 ± 0.0140 | γ2 | 0.9749 ± 0.0062 | 0.9478 ± 0.0117 |
| Card | α1 | 0.8624 ± 0.0099 | 0.8641 ± 0.0140 | α2 | 0.8779 ± 0.0160 | 0.8591 ± 0.0174 | β1 | 0.8561 ± 0.0502 | 0.8524 ± 0.0527 | β2 | 0.8814 ± 0.0137 | 0.8561 ± 0.0153 | γ1 | 0.8740 ± 0.0134 | 0.8596 ± 0.0197 | γ2 | 0.8879 ± 0.0120 | 0.8535 ± 0.0166 |
| Diabetes | α1 | 0.7506 ± 0.0231 | 0.7457 ± 0.0207 | α2 | 0.7843 ± 0.0156 | 0.7476 ± 0.0224 | β1 | 0.7570 ± 0.0151 | 0.7490 ± 0.0215 | β2 | 0.7780 ± 0.0143 | 0.7477 ± 0.0126 | γ1 | 0.7810 ± 0.0153 | 0.7370 ± 0.0152 | γ2 | 0.7902 ± 0.0134 | 0.7389 ± 0.0205 |
| Fertility | α1 | 0.8988 ± 0.0256 | 0.8218 ± 0.0616 | α2 | 0.9297 ± 0.0204 | 0.8170 ± 0.0551 | β1 | 0.9255 ± 0.0243 | 0.8309 ± 0.0459 | β2 | 0.9182 ± 0.0222 | 0.8279 ± 0.0467 | γ1 | 0.9455 ± 0.0199 | 0.8479 ± 0.0462 | γ2 | 0.9370 ± 0.0131 | 0.8236 ± 0.0484 |
|
|