Research Article
[Retracted] Identification of Diabetic Retinopathy through Machine Learning
Table 4
Precision results for optimum-path forest (OPF) (500 and 1000) and restricted Boltzmann machine (RBM) (500 and 1000).
| Repetitions | OPF-500 | OPF-1000 | RBM-500 | RBM-1000 |
| 0 | 0.7667 | 0.7667 | 0.868 | 0.868 | 1 | 0.8 | 0.8 | 0.789 | 0.895 | 2 | 0.7667 | 0.8 | 0.737 | 0.921 | 3 | 0.8 | 0.8 | 0.789 | 0.895 | 4 | 0.8333 | 0.8 | 0.711 | 0.921 | 5 | 0.7333 | 0.7333 | 0.763 | 0.868 | 6 | 0.8 | 0.7667 | 0.711 | 0.921 | 7 | 0.7667 | 0.7667 | 0.763 | 0.895 | 8 | 0.7667 | 0.7333 | 0.763 | 0.842 | 9 | 0.7667 | 0.7667 | 0.763 | 0.895 | 10 | 0.8 | 0.8 | 0.789 | 0.895 | 11 | 0.8667 | 0.7667 | 0.737 | 0.947 | 12 | 0.7667 | 0.7333 | 0.711 | 0.895 | 13 | 0.8 | 0.7667 | 0.737 | 0.895 | 14 | 0.6667 | 0.7667 | 0.816 | 0.868 | Average | 78 | 77.11 | 76.31 | 89.47 | Standard deviation | 4.51 | 2.48 | 4.31 | 2.64 |
|
|
Values in bold indicate the best results (mean and standard deviation) by the Wilcoxon test, with 5% significance.
|