Table 3 Antibiotic susceptibility classifier performance.
From: Personalized antibiograms for machine learning driven antibiotic selection
Antibiotic selection | Best model class | Prevalence | Average precision | AUROC |
|---|---|---|---|---|
Vancomycin | Gradient Boosted Tree | 0.23 | 0.46 [0.40, 0.52] | 0.72 [0.68, 0.75] |
Ampicillin | Gradient Boosted Tree | 0.43 | 0.54 [0.49, 0.58] | 0.62 [0.59, 0.65] |
Cefazolin | Gradient Boosted Tree | 0.59 | 0.72 [0.68, 0.76] | 0.67 [0.64, 0.70] |
Ciprofloxacin | Random Forest | 0.63 | 0.73 [0.70, 0.76] | 0.61 [0.58, 0.64] |
Ceftriaxone | Gradient Boosted Tree | 0.66 | 0.79 [0.77, 0.82] | 0.69 [0.66, 0.72] |
Cefepime | Random Forest | 0.80 | 0.87 [0.84, 0.89] | 0.65 [0.61, 0.69] |
Vancomycin + Ceftriaxone | Gradient Boosted Tree | 0.81 | 0.87 [0.84, 0.89] | 0.67 [0.63, 0.71] |
Meropenem | Gradient Boosted Tree | 0.82 | 0.90 [0.88, 0.92] | 0.69 [0.65, 0.72] |
Pip-Tazo | Random Forest | 0.90 | 0.94 [0.92, 0.95] | 0.64 [0.59, 0.69] |
Vancomycin + Pip-Tazo | Random Forest | 0.96 | 0.98 [0.97, 0.99] | 0.70 [0.62, 0.77] |
Vancomycin + Cefepime | Random Forest | 0.97 | 0.98 [0.98, 0.99] | 0.70 [0.62, 0.78] |
Vancomycin + Meropenem | Gradient Boosted Tree | 0.98 | 0.99 [0.99, 0.99] | 0.73 [0.65, 0.81] |
