From: miRDM-rfGA: Genetic algorithm-based identification of a miRNA set for detecting type 2 diabetes
GA settings | Setting 1 | Setting 2 | Setting 3 | Setting 4 | Setting 5 | Setting 6 |
---|---|---|---|---|---|---|
The number of selected features in the initial individuals (N) | 3 | 6 | 9 | 12 | 9 | 12 |
The optimal number of selected features (b) | 3 | 3 | 3 | 3 | 3 | 3 |
Population | 1000 | 1000 | 1000 | 1000 | 1000 | 1000 |
Generation (G) | 200 | 200 | 200 | 200 | 500 | 300 |
Crossover rate | 0.8 | 0.8 | 0.8 | 0.8 | 0.8 | 0.8 |
Mutation rate | 0.003 | 0.003 | 0.003 | 0.003 | 0.003 | 0.003 |
Penalty weight (W) | 20 | 15 | 10 | 7 | 10 | 7 |
Generation consisting of the best individual | 168 | 31 | 58 | 139 | 463 | 44 |
Selected miRNAs | hsa-miR-29b-1-5p, hsa-miR-6738-3p, and hsa-miR-125b-2-3p | hsa-miR-494-3p, hsa-miR-668-3p, and hsa-miR-29b-1-5p | hsa-miR-222-5p, hsa-miR-671-5p, and hsa-miR-1307-3p | hsa-miR-494-3p, hsa-let-7b-5p, and hsa-miR-29b-1-5p | hsa-let-7b-5p, hsa-miR-125b-5p, and hsa-miR-7-5p | hsa-miR-7-5p, hsa-miR-92b-3p, and hsa-let-7b-5p |
Fold for cross-validation of test data | 3 | 3 | 3 | 3 | 3 | 3 |
Mean AUROC score by threefold cross-validation in test set and standard deviation | 0.86 ± 0.09 | 0.89 ± 0.08 | 0.87 ± 0.05 | 0.89 ± 0.06 | 0.92 ± 0.04 | 0.90 ± 0.06 |