Table 3 Training accuracy v.s.testing accuracy on heart dataset. Accuracy indicates testing accuracy. Accuracy indicates test accuracy. At 12 trainings, it can be considered that learning has just started, and at 72,120 trainings, it can be considered that model building is in the final stage
From: Performance of quantum kernel on initial learning process
Training size | Classic | Y feature map | Z feature map | |||
|---|---|---|---|---|---|---|
Training Accuracy | Testing Accuracy | Training Accuracy | Testing Accuracy | Training Accuracy | Testing Accuracy | |
12 | 0.667 | 0.550 | 1.000 | 0.522 | 1.000 | 0.523 |
72 | 0.889 | 0.775 | 0.986 | 0.792 | 0.986 | 0.814 |
200 | 0.905 | 0.806 | 0.980 | 0.835 | 0.980 | 0.845 |