Table 5.
Regression performance comparison using estimated and actual GEX values on Dataset1 and Dataset2.
| Method | Using estimated GEX | Using actual GEX | |||
|---|---|---|---|---|---|
| MAE | MSE | MAE | MSE | ||
| Dataset1 | REP-ElasticNet | 1.131 | 2.423 | 0.919 | 1.504 |
| REP-KNN | 1.087 | 2.281 | 0.839 | 1.305 | |
| REP-RandomForest | 0.893 | 1.513 | 0.809 | 1.098 | |
| REP-SVR | 1.105 | 2.049 | 0.727 | 0.891 | |
| Dataset2 | REP-ElasticNet | 0.918 | 1.335 | 0.934 | 1.237 |
| REP-KNN | 0.852 | 1.176 | 0.658 | 0.780 | |
| REP-RandomForest | 0.813 | 1.148 | 0.930 | 1.265 | |
| REP-SVR | 0.84 | 1.144 | 0.660 | 0.792 | |
Columns corresponding to “estimated GEX” represent analysis in which in the test set, only the GEX values at the initial time point were used and the GEX of later time points were estimated. Columns corresponding to “actual GEX” represent analysis in which in the test set, GEX values of all time points were used.