Table 1.
Summary of the performance and number of attributes used for the different models studied in this paper.
| Model | Number of parameters | c-indexes | |
|---|---|---|---|
| Discovery | Validation | ||
| Cox | 4 | 0.735 | 0.744 |
| Best linear (OLPM) | 4 | 0.771 | 0.817 |
| NN with CV | 4 | 0.791 | 0.825 |
| NN | 4 | 0.740 | 0.751 |
| RFF_SVM | 4 | 0.747 | 0.783 |
| libSVM | 4 | 0.739 | 0.756 |
| RT | 4 | 0.696 | 0.681 |
| NN | 44 | 0.794 | 0.746 |
| RFF_SVM | 44 | 0.801 | 0.766 |
| libSVM | 44 | 0.752 | 0.700 |
| RT | 44 | 0.741 | 0.630 |
Cox, best optimized linear prognosis model and machine learning-based approaches are included. Results are listed for both the discovery and validation cohorts. C-indexes in the validation group over 0.8 are boldfaced.