1

This is similar to Ratios in Regression, aka Questions on Kronmal, although the question there is about linear regression. The advice in the linked paper there, which you should have a look at, is to use as regressor variables the numerator, inverse denominator, and their interaction (that is, the ratio). You could at least try that! (and tell us how it went)...


Only top voted, non community-wiki answers of a minimum length are eligible