Matrix reduction from large matrix data.

I have an input training data of dimension 10,451 by 41.

How can i form a 3 by 3 matrix from each of the row (1 by 41)???

[Durganshu]

Please elaborate on what is the basis of your matrix reduction. Are you trying to implement something like PCA?

[KSSV]

As you have 41 elements in a row.....it is tough to divide them into 3*3 matrices as you will be left with some elements left or append extra zeros/ nans to the array.

A = rand(1,41) ;
B = reshape(A(1:36),3,3,[]) ; % 36 because it is divisble by 9 
Anew = [A NaN(1,4)] ; 
Bnew = reshape(A,3,3,[]) ;   % append nans to make A of length 45