I am not very familiar with cross-correlation analysis. I am analysing a large number of "paired" time series. For each pair, I am doing a ccf in R. I read in other posts that the horizontal line given by R indicates the significant values. How are they tested to be significant? Because I have to do this for hundreds of pairs, is there a way of grabbing the significant number from the R output?

> print(ccf(x,y))
Autocorrelations of series ‘X’, by lag                                 
-6     -5     -4     -3     -2     -1      0      1      2      3      4                                                                    
-0.242 -0.090  0.057  0.197  0.466  0.699  0.896  0.436  0.221 -0.018 -0.116 

Related to the R output, I noticed that these values are called autocorrelation coefficients? But they describe the correlation of a lagged A series with series B right? Does anybody have any more insight into why it's called autocorrelation?


1 Answer 1

  1. Yes, the values you get describe the correlation of lagged A series (or x in your code) with series B (or y in your code).

  2. It is called autocorrelation, because cross-correlations of series A, and B, come from auto-correlation of multivariate time series $(A,B)$. For univariate time series auto-correlation is scalar, for multivariate it is matrix. It is still auto, meaning that it is correlation with itself.

  3. If you only need the values, use plot = FALSE in the call to ccf.

  • 1
    $\begingroup$ Thanks for your comment, very enlightening! so say in this case that I take the highest value: 0.896. How do I see that it's above the horizontal line that indicates significance? If these autocorrelation values are the correlation coefficient (r), what is the significance value? If the lag is over the line (significant) reported in my stats software, does this mean p < 0.05? $\endgroup$
    – dorien
    Jan 21, 2016 at 14:57

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.