EViews.com

Hello,

I'm familiar with the syntax for displaying sample correlations in eviews, i.e. : Question #1: instead of an output matrix of sample correlations, I want to code a vector that displays only the correlations between, say, x1 & x2 & x3, (omitting the entry that is correlation with itself). Perhaps this done through referencing the exact element in the correlation matrix or some other related syntax?

Question #2: Does eviews have a dedicated syntax for autocorrelation (first order, 2nd order, ect) for a series called x1? In this case, viewing any i-th order autocorrelation for series x1 in scalar form would work for me.

You'll have to extract the non-diagonal elements from the correlation matrix - we don't have a way to just calculate the cross-correlations without the diagonals.

series.correl shows the autocorrelations.

Thanks for clearing the first part up.

The autocorrleation portion though, I was more after a syntax to reference the autocorrelation by its order, rather than call the user interface.

for example if I want to view the 3rd order auto correlation of series "x1" in scalar form, what kind of code would eviews need?

series.correl. The full syntax is explained in the Object Reference.

Looking through version 8's command reference, I did find a segment covering autocorrelation by order, like:
however it doesn't seem to work for viewing the autocorrelation entry as a scalar...
returns the error: "correl(3) is not a member or procedure of x1"

ideally, I'd like to create a scalar that contains the 3rd degree autocorrelation value for the series.

any ideas?

---UPDATE--- 06/16/2015

The object reference does explain the syntax, but it basically just said series_name.correl(n, options).

The closest I got to referencing the autocorrelation by order within a scalar was by freezing into a table.


I have two questions:

1. When I call the function x1.correl it displays all lags up to n, in the above case 1 to 3. Is there any way to call only n, and no lags before?

2. Is it possible to freeze into a scalar instead of a table? If not, what is the best way to convert a table element into a scalar?

Please advise

Thanks again