Table 2 Performance (AUC) results of six machine learning methods
From: Computational determination of hERG-related cardiotoxicity of drug candidates
Algorithm | Optimal number of features | AUC |
|---|---|---|
Linear regression | 40 | 0.747 |
Logistic regression | 350 | 0.764 |
Ridge regression | 400 | 0.774 |
Neural network | 1400 | 0.764 |
Naïve Bayes | 40 | 0.687 |
Random forest | 120 | 0.709 |