1
$\begingroup$

I have created a ridge and lasso regression model in R. From my understanding, the coefficients are interpreted differently from logistic regression. For instance, in logistic regression you may say males have a 4 times greater odds of developing a certain condition than females (assuming the coefficient was 4 and the predictor variable was males, with females as the reference). A significance value is also provided.

How do you interpret the coefficients in ridge and lasso regression, especially considering there is no p value?

Improve this question
$\endgroup$
1
  • 1
    $\begingroup$ Te lack of p-values and confidence intervals are a separate issue from what I address here, but for the point estimates, I argue that they should be interpreted the usual way. $\endgroup$
    – Dave
    Commented Dec 20, 2021 at 11:37

1 Answer 1

Reset to default
2
$\begingroup$

It is not generally a good idea to interpret the coefficients in a elastic net model. The coefficients are penalized and are not suitable for inference about the true value of the parameter. These models are typically used for prediction, and should be evaluated based on the accuracy of their predictions.

Improve this answer
$\endgroup$
2
  • 1
    $\begingroup$ Could you elaborate on why penalising the coefficients makes the model unsuitable for inference? $\endgroup$
    – Accidental Statistician
    Commented Aug 21, 2021 at 23:53
  • 2
    $\begingroup$ @AccidentalStatistician this is an old question/answer and the answered has only ever answered this one question. You may get better responses by asking your own question. $\endgroup$
    – Robert Long
    Commented Aug 22, 2021 at 7:54

Your Answer

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

Not the answer you're looking for? Browse other questions tagged or ask your own question.