Table 1.
List of hyperparameter values.
Fig 1.
Accuracy vs. number of SNP for SVM (left) and kNN (right).
Table 2.
Comparison of feature selection methods.
Fig 2.
Feature importance score of top 20 SNPs.
Table 3.
Characteristics of top five SNPs selected by RF based on importance score [21].
Table 4.
Comparison of performance between RF-SVM and RF-kNN to the baseline SVM and kNN models.
The selected SNPs for Baseline SVM and kNN refers to the 19 asthma-associated SNPs.
Fig 3.
Performance of the three ML models (RF-SVM, RF-kNN, MeanDiff-kNN).
Table 5.
Optimal hyperparameters determined for the models.
Fig 4.
ROC Curves of RF-SVM (a), RF-kNN (b), and MeanDiff-kNN (c).