Table 2 Analysis of average AUC values for 18 discriminant models with preprocessing.
Feature selection methods | Box-Cox | Max-Abs | Min–Max | Quantile | Yeo-Johnson | Z-score | ||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Training | Test | Training | Test | Training | Test | Training | Test | Training | Test | Training | Test | |
Logistic regression | 0.914 | 0.846 | 0.907 | 0.839 | 0.893 | 0.833 | 0.903 | 0.852 | 0.911 | 0.845 | 0.911 | 0.845 |
SVM | 0.913 | 0.837 | 0.981 | 0.823 | 0.909 | 0.824 | 0.906 | 0.827 | 0.906 | 0.829 | 0.906 | 0.829 |
Random forest | 0.965 | 0.775 | 0.942 | 0.801 | 0.941 | 0.803 | 0.973 | 0.735 | 0.956 | 0.786 | 0.953 | 0.780 |
