Table 4. Externally validated model discrimination for different predictor sets and modelling techniques: Validation cohort.
Predictor set | Model | ||
---|---|---|---|
CPH | RF | GBC | |
QA | 0.736 | 0.736 | 0.796 |
QA+ | 0.743 | 0.799 | 0.810 |
T | 0.788 | 0.810 | 0.826 |
Predictor set T and GBC modelling constantly perform better than their counterparts. The results conform to the pattern observed in internal cross-validation.
CPH, Cox proportional hazards; GBC, gradient boosting classifier; RF, random forest.