Skip to main content
1 of 3
Miguel M.
  • 558
  • 1
  • 4
  • 18

Does the "partial" version of the Cross Correlation Function exist? If so, how can it be computed?

I have been trying to look for the analogous version of the PACF for cross correlation, but there seems to be no such thing... why is this? Is it not possible to extrapolate the algorithm used to calculate the PACF to two time series? From my understanding, the CCF is merely the correlation between the lags and leads of two time series, as it is defined here: Wikipedia page about cross correlation

Am I missing something about the CCF? Because its computation is very simmilar to how the ACF is constructed... so the "PCCF" should be constructed simmilarly to the PACF?

Also, if anyone could provide some R code to compute the aforementioned "PCCF", it would be much appreciated... I found a "PCCF" function in Matlab, but I am uncertain if it is correct: Matlab CCF function

Miguel M.
  • 558
  • 1
  • 4
  • 18