Table 3. Accuracy of machine learning algorithms.
| Algorithm | AUC | Features used |
|---|---|---|
| Decision tree | 0.933 | 2/65 |
| Random forest | 0.952 | 9/65 |
| Support Vector Classification | 0.965 | 5/65 |
| Logistic Regression | 0.962 | 5/65 |
| Categorical lasso | 0.962 | 5/65 |
| Linear discriminant analysis | 0.964 | 5/65 |
Using mutual information feature selection methods. Categorical Lasso was implemented as a Logistic Regression model with l1 regularization and our Logistic Regression model utilized l2 regularization. Support Vector Classification was applied using the linear kernel.