0
$\begingroup$

Assume I have:
N = 100, 5 continues predictor variables and 1 continues response. Now I want to find a mapping between predictors and response using multiple linear regression. Lets assume I know that my 100 samples consist of x subgroups, where for any subgroup I could find a match with very low MSE but for the whole population my MSE would increases drastically.

Is there a standardized way to find out whether these subgroups exist?

Improve this question
$\endgroup$
2
  • $\begingroup$ You say you "know" the subgroups exist and then you say you want to "find out" if they do. Do you mean you want to find which subjects go into which cluster? Or something else? $\endgroup$
    – Peter Flom
    Commented May 20, 2014 at 10:17
  • $\begingroup$ Sorry i was not clear there, i want to assume i dont know subgroups exist and i want to find out whether they do or not. $\endgroup$
    – Susi_28
    Commented May 20, 2014 at 10:37

0

Reset to default

Your Answer

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

Browse other questions tagged or ask your own question.