How to calculate moving covariance in a matrix?

Hi guys, Hi Guys, I have got a matrix :378x9. I need to calculate the moving covariance with a window size of 120(starting from row one). Can somebody help me please? Thank you very much Andrea

[Image Analyst]

Covariance between what two random variables? What does each set of values represent and where do they come from?

[Andrea Finocchiaro]

Thank you for replying! Covariance between 9 random variables. The values represent the return of 9 stocks. I have 378 information for stock.Matrix 378x9.

[Image Analyst]

OK, so you're finding out the correlation coefficient between column 1 and col 2, col 1 and col 3, etc. But what do you mean by having a moving window? Do you want to not use the whole column, but just a subset/window of the columns, and then do it again with the window moved down until the whole column has eventually been looked at?

[Andrea Finocchiaro]

Yes,correct that is the moving window I mena. But why correlation?I am looking for the Covarinces

Answer 1

[the cyclist]

% Generate some pretend data. Use your real data here.
data = rand(378,9);

[M,N] = size(data);
window = 120;

numberCovarianceMatrices = M - window + 1;

covarianceMatrices = zeros(N,N,numberCovarianceMatrices);

for nc = 1:numberCovarianceMatrices
    covarianceMatrices(:,:,nc) = cov(data(nc:(nc+window-1),:));
end