My plots of conditional weighted residuals (CWRES) plotted against time show some sort of time trend (image attached). The response variable is on a Box_cox scale. How could I solve this problem ? enter image description here

  • 1
    $\begingroup$ "Box-Cox scale" isn't precise enough to indicate what you did, but some transformation of the response is implied. But Box-Cox does not purport to do anything to remove time trend. The time trend here seems slight. Whether (a) it is a side-effect of the transformation (b) you should modify your model to match the time trend (c) you can just ignore it -- or some combination -- is not something on which we can easily advise on this little information. $\endgroup$ – Nick Cox Apr 7 '15 at 10:21
  • $\begingroup$ In addition, not that appearance of trend can be sensitive to how and how much you smooth. $\endgroup$ – Nick Cox Apr 7 '15 at 10:26

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.