Convert for loop to vector multiplication.

Hello Friends,

I have the following:

A = [1 2 3;
     4 5 6;
     7 8 9];

B = [10 11 12;
     13 14 15];

[N1, D1] = size(A);
[N2, D2] = size(B);

for i = 1:N2    
    diffs = bsxfun(@minus, A, B(i, :));    
    dists(:, i) = sqrt(sum(diffs.^2, 2));
end

How can I replace the 2nd line in for loop in order to write it in a vector form? I have tried something like this:

for i = 1:N2    
  diffs = bsxfun(@minus, A, B(i, :)); 
  dists = sqrt(diffs.*diffs);
end

But the two results are different. I am trying to achieve the following:

distance = sqrt((A-B).*(A-B))

which should be the Euclidean distance. I want to keep it as presented above without changing format to compute distance or using repmat, etc.

Answer 1

[Walter Roberson]

dists = sqrt(diffs.*diffs);

is not going to be doing any sum. You might be thinking of the * operator instead of the .* operator.

You could also look at dot() and norm()

[hello_world]

Can you elaborate please with a work example using * so that the two results be the same.