How to calculate averages for each rows from a specific column to a specific column in a matrix?

So, I have a large matrix (rows = 1000; columns = 1000) and I have to compute the averages of each rows, but not the whole, only 100 columns, each step.

Let me explain:

There is a given matrix.

A = 
1  2  3  4  5  6
7  8  9  10  11  12
13  14  15  16  17  18
19  20  21  22  23  24
25  26  27  28  29  30
31  32  33  34  35  36

In this case, I have to take the average of 2 columns (step by step) in each rows:

(1+2)/2   (3+4)/2   (5+6)/2
(7+8)/2   (9+10)/2  (11+12)/2
(13+14)/2 (15+16)/2 (17+18)/2  
...

The solution is a new matrix:

B = 
1,5  3,5  5,5
7,5  9,5  11,5
13,5  15,5  17,5
19,5  21,5  23,5
25,5  27,5  29,5
31,5  33,5  35,5

How can I code this to make the same results in bigger sizes (to calculate the average in each row by columns 1-100; 101-200; 201-300; etc...)?

Any idea is highly welcomend!

Thank you!

Answer 1

[the cyclist]

Here is one way, using the movmean function.

A = ...
[1  2  3  4  5  6
7  8  9  10  11  12
13  14  15  16  17  18
19  20  21  22  23  24
25  26  27  28  29  30
31  32  33  34  35  36];

NC = 2; % This is the number of columns you want to average over.

tmp = movmean(A,NC,2); % Note that the "2" here is dimension to average over, NOT the number of columns

B = tmp(:,NC:NC:end);

[Roland Baczur]

Thanks, it works!

Answer 2

[Jan]

A = [1  2  3  4  5  6; ...
     7  8  9  10  11  12; ...
     13  14  15  16  17  18; ...
     19  20  21  22  23  24; ...
     25  26  27  28  29  30; ...
     31  32  33  34  35  36];
sA = size(A);
AA = reshape(A, [sA(1), 2, sA(2)/2]);
Am  = squeeze(sum(AA, 2) / 2);

In the general case the size of A might not be a multiple of the block width. Then see FEX: BlockMean: Create the mean over N columns:

S  = size(A);
Sf = S(2) - mod(S(2), N);  % Crop trailing columns
Nv = Sf / N;
AA = reshape(A(:, 1:Sf), S(1), N, Nv);
Am  = sum(AA, 2) .* (1.0 / N);
Am  = reshape(Am, S(1), Nv);

[Roland Baczur]

Thank you too, it is also working (right now, I'm using the movemean function).