About maximum value in matrix

I have a problem regarding finding a maximum value from my matrix (largest element in the row)

I have matrix A shown bellow

I know that I can use

M=max(A,[],2)

Then I will get matrix M such that

But the answer I want is not that, what I want is that . The one that I want is

The answer that I want is the maximum value of each row ignoring the negativity of each value like shown in the matrix above.

I hope my question is clear enough.

Answer 1

[Stephen23]

This can be done using sub2ind:

A = [2,4,-1,0,-1;1,3,12,-3,2;-12,2,4,0,-1;2,4,0,-5,-2;3,-4,-1,0,8]
[~,C] = max(abs(A),[],2); % C = column indices
S = size(A);
R = 1:S(1);               % R = row indices
X = sub2ind(S,R(:),C(:)); % X = linear indices
Z = A(X)

Giving

Z =
    4
   12
  -12
   -5
    8

It would be nice if max etc. could return the linear indices, but that is what we have to work with. Note that you can also calculate the linear indices yourself, without sub2ind:

X = S(1)*(C-1)+(1:S(2)).'

[Mohamad Agung Prawira Negara]

Thanks for the answer, it works for me

[Stephen23]

Another possibility:

>> A = [2,4,-1,0,-1;1,3,12,-3,2;-12,2,4,0,-1;2,4,0,-5,-2;3,-4,-1,0,8];
>> [C,X] = max(max(cat(9,A,-A),[],2),[],9);
>> C .* (3-2*X)
ans =
    4
   12
  -12
   -5
    8

Answer 2

[Fangjun Jiang]

use abs() and note that the second return of max() gives the indices. See if you can figure it out.

[Mohamad Agung Prawira Negara]

thanks for the hint