How to compare some specific entries of a matrix, if I have their indices stored in 2 files?

I have 2 files, that have numbers that represent rows and columns respectively, out of which I want to pick a pair at a time, and compare that corresponding element from a given matrix. For example, if I have matrices A and B, such that:-

A= [1,5,9] and B= [2,6,11]

Then I want to compare the (1,2), (5,6) and (9,11) elements of a matrix C, and return the indices of the largest value.

Originally, my matrix A and B are 19007x1 in size. And my matrix C is 5601x5601 in size.

Answer 1

[José-Luis]

 idx = sub2ind(size(C), A, B)
 result = max(C(idx));

[Aditya Wahi]

Will you please explain what exactly is happening in these 2 lines? And what is the result returning. In my case,

 result= 4.314598144531250e+03

What is it?

[José-Luis]

I misread your question, it's returning the maximum value, not the index you asked for.

 idx = sub2ind(size(C), A, B);
 result_idx = find(C == max(C(idx)));
 res_a = A(result_idx);
 res_b = B(result_idx);

[Aditya Wahi]

I ran this code, but it is showing the following error:

Attempted to access row(6.71872e+06); index out of bounds because numel(row)=19007.

[Aditya Wahi]

My c matrix is called v_net, A is row and B is col

[José-Luis]

Sorry, I tried to wing it without starting up Matlab.

 A =  [ 1 , 2 , 7 ];
 B = [ 5 , 4 , 8];
 C = rand(10);

 idx = sub2ind(size(C),A,B);
 [result_x, result_y] = find(C == max(C(idx)));

[Guillaume]

The problem is with the find(C == max(C(idx))), you can't use find on the full C matrix, you have to do it on the subset indexed by idx, so:

find(C(idx) == max(C(idx)))

edit, after José-Luis edit while I was writing my comment: while the new code may provide correct results most of the time, this is still wrong. As it searches the whole C matrix, not just the subset provided by A and B. Try with example:

A = 1;
B = 1;
C = [0 0; 0 0];

[José-Luis]

True, didn't think of that. Yet another edit.

A =  [ 1 , 2 , 7 ];
B = [ 5 , 4 , 8];
C = rand(10);
idx = sub2ind(size(C),A,B);
[idx_x, idx_y] = find(C(idx) == max(C(idx)));
x = A(idx_x);
y = B(idx_y);

Answer 2

[Guillaume]

As per José-Luis' answer, you have to use sub2ind to convert your 2D indexing from A and B into linear indices. This would work:

index = sub2ind(size(C), A, B);
[~, row] = max(C(index));

row is the row in A and B where the maximum is found. If there are several locations where it is found, then you only get the first one. If you want all of them:

Csearch = C(sub2ind(size(C), A, B));
rows = find(Csearch == max(Csearch));