9
$\begingroup$

I wish to test the cross-correlation of two time series to measure how "instantaneously related" my time series are. However, I am not quite sure if I should difference my time series first, and make them stationary? Because they obviously aren't:

enter image description here

enter image description here

And the cross-correlation I get is very different if I difference my time series, or not.

Cross-correlation cff() on original non-differenced data:

enter image description here

Cross-correlation cff() on differenced data with diff():

enter image description here

Which one is right? Am I supposed to use cff() on stationary data only? Thank you for the help.

$\endgroup$

1 Answer 1

3
$\begingroup$

why don't you post your data and I will try and help you. Box and Jenkins suggested pre-filtering where the differencing operator identified as part of the ARIMA process was used as part of the filter to identify model form via the resultant cross-correlation . See https://web.archive.org/web/20160216193539/https://onlinecourses.science.psu.edu/stat510/node/75/ AND http://www.math.cts.nthu.edu.tw/download.php?filename=569_fe0ff1a2.pdf&dir=publish&title=Ruey+S.+Tsay-Lec1 AND http://autobox.com/cms/images/dllupdate/TFFLOW.png . However if both X and Y are I(1) .. it is still possible that Y and X can related without any differencing at all which suggests that alternative approaches to incorporating differencing in the TF model or not be on the table.

$\endgroup$

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.