This tutorial aims to show a broad audience of applied statisticians and epidemiologists that some ‘non-standard’ shrinkage methods may be very useful for typical epidemiological studies that target multivariable prognostic models.
In this tutorial “Think before you shrink” you’ll learn more about shrinkage. Prof. Van de Wiel will discuss ridge shrinkage and elaborate on how dedicated shrinkage improves your results for simple regression problems, and why may it be useful in challenging low-dimensional regression settings.
Preliminary knowledge: Regression models, ordinary least squares. The concept of shrinkage will be (re-)explained, although it helps when one knows a bit about techniques like ridge and lasso. No specific knowledge on Bayesian statistics is required. Formula density will be low. Based on joint work with Jeroen Hoogland, Koos Zwinderman, Martijn Heymans, Gwenaël Leday (WUR), Erik van Zwet (LUMC).
| Start date: | |
|---|---|
| End date: | |
| Time: | |
| End time: | |
| Location: | VU, MF building, A415 |
| Costs: | Free |
|---|---|
| Register now |
| Organisor |
APH Methodology |
|---|---|
| Contact | aph.methodology@amsterdamumc.nl |