Quantitative Methods 2017

Previous Page

Next Page

Page Background

How good is the Model?

1. Discrimination (AUC, C index…)

a)

Training dataset

b)

External validation dataset

2. Calibration on external dataset

3. Interpretability

4. “Tripod” evaluation

5. Usable in distributed setting

6. Comparison with gold standard

7. Clinical usefulness

8. …