1
$\begingroup$

Is penalized logistic regression only appropriate when the likelihood of the outcome is very low, or when the number of variables exceeds the number of observations?

Wondering if it can be a tool to use generally in order to shrink variables that "contribute" less to the observed effect in epidemiology.

$\endgroup$

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.