Table 4.
Results of classification incorporating Shufflenet deep features applied on the HAM10000 dataset.
| Classifiers | Accuracy (%) | Recall (%) | Precision (%) | F1 Score | FNR | Time (sec) |
|---|---|---|---|---|---|---|
| Narrow NN | 96.5 | 96.82 | 96.75 | 96.68 | 3.18 | 197.3 |
| Medium NN | 97.3 | 97.55 | 97.51 | 97.52 | 2.45 | 116.01 |
| Wide NN | 97.6 | 97.78 | 97.72 | 97.74 | 2.22 | 130.81 |
| Bilayered NN | 96.6 | 96.84 | 96.81 | 96.82 | 3.16 | 284.3 |
| Quadratic SVM | 98.1 | 98.3 | 98.34 | 98.3 | 1.7 | 546.39 |
| Cubic SVM | 98.5 | 98.62 | 98.64 | 98.62 | 1.38 | 585.07 |
| Coarse Gaussian SVM | 96.5 | 96.71 | 96.77 | 96.72 | 3.29 | 986.83 |
| Fine KNN | 98.9 | 99.15 | 99 | 99.06 | 0.85 | 920.22 |
| Ensemble Bagged Tree | 95.0 | 95.13 | 95.34 | 95.32 | 4.87 | 209.95 |
| Fine Tree | 69.3 | 68.62 | 71.21 | 69.92 | 31.38 | 63.97 |
Bold show the significant value.