Table 3 Performances of selected classifiers to predict a patient with high opioid dosing
Classifier | Group difference (P-uncorrected)* | No. of genes | Sensitivity (%) | Specificity (%) | PPV (%) | Accuracy (%) | No. of 7 independent samples correct |
|---|---|---|---|---|---|---|---|
Chr6:154360797-SNV | P=0.589 | 1 | 13.9 | 87 | 51 | 50.6±5 | 0 |
Chr6:154451812-SNV | 0.00049 | 1 | 2.9 | 100 | 29 | 51.5±2 | 0 |
Chr8:54147491-SNV | 0.000278 | 1 | 10.5 | 92.2 | 52 | 51.4±5 | 4 |
Regression | — | 6 | 67.7 | 55.7 | 60 | 61.7±6 | 4 |
CART | — | 12 | 92.6 | 56.1 | 68 | 74.4±4 | 3 |
Clustering+kNN classifier | — | 152 | 75.3 | 73.5 | 74 | 74.4±7 | 5 |
ABC analysis+clustering+kNN classifier | — | 34 | 93.7 | 67.4 | 74 | 80.6±4 | 4 |