It is well known that if two linear regression predictors highly correlate, it is bad for our model, but which correlation is considered to be big? Is it 0.5,0.6,0.8,0.9..? I have tried to find out about this on the internet, but I didn't manage to find exact value.

  • 2
    $\begingroup$ High correlation can be a problem. There in not a „exact“ threshold value to judge this. The variance inflation factor is a widely used method to check for (too) high multicollinearity datascience.stackexchange.com/a/64415/71442 $\endgroup$
    – Peter
    Jan 13 at 18:14
  • $\begingroup$ I posted an answer on Cross Validated last year that is likely to be of interest, since I get a sense that you want to drop a variable if you find two to be correlated. $\endgroup$
    – Dave
    Jan 13 at 21:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.