Table B1.
Execution times for feature selection and model training algorithms (including parameter optimisation) obtained for a binary classification problem. Training times were obtained with a subset of 50 mRMR features and 4,072 training instances. Simulations were performed using an intel core i5-2520 M CPU 2.50 GHz, 4 GB RAM
| Feature selection | Execution time (ms) |
|---|---|
| FCBF | 18,922 |
| IG | 787,152 |
| Relief | 2,612,942 |
| Chi-square | 4,671 |
| SU | 2,008 |
| CFS | 2,884,391 |
| mRMR |
21,652 |
| Prediction model |
Training time (ms) |
| DT (mRMR) | 30,585 |
| NB (mRMR) | 141,530 |
| Logit (mRMR) | 12,341 |
| SVM (mRMR) | 59,465 |
| DNN (mRMR) | 85,770 |