Averaging data every N rows

hi all, I have a matrix in dim of 8761*5 I want to average the data every 12 rows, starting from the 7th row to 19th and so on. the last 6 rows should be with the first 6 rows. the final dim should be 731*5

thanks for help.

[Image Analyst]

If you ignore the first 6 rows, then you have 8755 left. Taking groups of 12, you'd have 729.583333333333 sets of 12 because it's not a multiple of 12. So how do you get 731?

And we have no idea what "the last 6 rows should be with the first 6 rows." means. Does it mean group rows 1-6 (first 6) together with rows 8756-8761 (the last 6) when taking the mean?

[Lilya]

That's the problem. Yes it is, the first 6 rows should be with the last 6 rows It's a half-daily data

Answer 1

[Image Analyst]

Try this, if you have the Image Processing Toolbox.

data = randi(9, 8772, 5); % Sample data.
% Start with the 7th row by doing an upward shift of 6 rows.
newData = circshift(data, -6, 1);

% Block process the matrix to give the mean of the pixels in a 12 row by 1 column block.
% The output will have 5 columns and 8772/12 = 731 rows.
meanFilterFunction = @(theBlockStructure) mean(theBlockStructure.data(:));
blockSize = [12, 1];
% Now compute the block means.
blockMeans = blockproc(newData, blockSize, meanFilterFunction)
[rows, columns] = size(blockMeans) % Just to double check the size.

[Lilya]

Thank you

Answer 2

[Image Analyst]

Try

data = [data(6:end-6);data(1:6,:)];

or use circshift().