EViews.com

I have a time series with several thousand observations and want to examine it for correlation within rolling twenty-day periods (so that, for example, if I had 4000 observations, I would have 3,980 different correlograms to examine). I would then like to slice and dice the correlation data into new series comprised of all the 1st-lag correlation, and then all the 2nd-lag correlation, then the 3rd, etc. Obviously, this is impractical to do by changing the sample size over and over. I am new to Eviews programming (the majority of my programming experience is in Java). What's the best way to handle this? Thanks in advance.

(Note - previously posted in wrong forum; earlier post has been deleted.)

The simplest way *is* to change the sample over and over again, but using program loops. Have you looked at the documentation on loops. You might also want to look at the example programs describing how to do rolling regression (or search the forums). The ideas there should transfer directly.