Calculate arithmetic average based on block numbers

Hi there,

This is a 2D matlab cell that I have (SEE ATTACHED IMAGE)

I want to average block numbers 1,2,3,4 and add into a new matrix (say A), and then average of 5,6,7,8 in 2nd row of matrix A, and then average of 9,10,11,12... in the 3rd row, etc.

Any help will be greatly appreciated.

Thanks, Akul

Answer 1

[dpb]

Don't use figures, attach .mat files or paste short sections of text for data...nothing anybody can do with a figure except look.

But, in your case it's easy enough using the fact that ML internal storage is column-major and the builtin functions ability to operate on arrays--

X=xlsread('yourfile.xls');          % get the data into an array
A=mean(reshape(X,4,[]));            % average by column sets of four variables
A=reshape(X,[],size(X,2));          % rearrange means by original columns

The above averages the block numbers as well; you can either

1. Choose to use that average, too,
2. Select X(:,2:end) instead and ignore column (changing size to match), or
3. Ignore or overwrite the block column after done.

I'll use option 3) here--

A(:,1)=(1:size(A,1)).';        % write the new number of average block

[Akul Satish]

That's great! Thank you so much :)

[Akul Satish]

However, I had to change the second line to:

A=reshape(A,[],size(X,2));

[dpb]

I see I missed second closing parenthesis...it needs must be

A=mean(reshape(A,4,[]));

The rearrangement to four rows is needed to average four elements using the default orientation of mean to operate by column.

Answer 2

[Image Analyst]

This should never have been a cell to begin with. So use cell2mat() to convert to a matrix. Then loop over columns reshaping to 4 wide, and call mean:

A = cell2mat(cellArray);
%A = randi(4, 12, 13)
[rows, columns] = size(A);
meanValues = zeros(rows/4, 13);
for col = 1 : columns
  thisCol = A(:, col);  % Must be a multiple of 4!!!
  reshaped = reshape(thisCol, 4, [])
  meanValues(:, col) = mean(reshaped, 1);
end
meanValues