Results for multicollinearity on the top 10 features using the variance inflation factor (VIF). A maximum VIF of 10 is chosen as the threshold and removing features with VIF greater than the threshold is said to reduce the multicollinearity. We see from table 1 that the features P_F269 and P_F289 have VIF >10, meaning these predictors are highly correlated with the remaining features in the model
| Top 10 features used for the prediction of RECIST response | Top 10 features used for the prediction of irRECIST response | ||
| Feature | VIF | Feature | VIF |
| P_F7 | 1.36 | P_F89 | 1.57 |
| P_F84 | 1.49 | P_F187 | 1.21 |
| P_F269 | 15.12 | P_F269 | 12.78 |
| P_F289 | 15.40 | P_F289 | 12.01 |
| PV_F35 | 5.38 | PV_F24 | 1.38 |
| PV_F89 | 5.22 | PV_F81 | 1.27 |
| PV_F109 | 9.19 | PV_F270 | 1.35 |
| Ar_F152 | 1.09 | Ar_F152 | 1.35 |
| Ar_F154 | 1.16 | ArV_F36 | 4.13 |
| Ar_F238 | 1.17 | ArV_F247 | 3.74 |
RECIST, Response Evaluation Criteria in Solid Tumors
irRECIST, immune-related Response Evaluation Criteria in Solid Tumors