How do I change a whole row/column to zero in a matrix relating to a 'greater than' condition.

I have a random network and I would like to make the rows which have the largest sum equal to zero.

A=createsrandomnetwork(100,5)

U=sum(A)

Say the 1st and 50th number in U are the largest, I would want to make that number row and column equal to zero. I know that I could do it by noting down these numbers and then doing this:

A(1,:)=0

A(50,:)=0

A(:,1)=0

A(:,50)=0

But is there an easier, more automatic way where I could say if a value in U is greater than 5 then change the corresponding column/row to zero?

Answer 1

[Guillaume]

To set the rows/columns where the maximum sum is found:

maxsum = max(U);  %I'm assuming that U is a vector here
A(U == maxsum, :) = 0;
A(:, U == maxsum) = 0;

To set rows/columns where U is greater than a value

A(U > 5, :) = 0;
A(:, U > 5) = 0;

[KJones]

Thank you, this was very helpful. I was also wondering if I wanted to make the 4 highest U value rows go to zero, how would I go about doing that?

[Guillaume]

sort the U vector, the first four values of the 2nd output of sort are the indices of the rows you want to zero:

[~, idx] = sort(U, 'descend');
A(idx(1:4), :) = 0;

Answer 2

[Alexandra Harkai]

> But is there an easier, more automatic way where I could say if a value in U is greater than 5 then change the corresponding column/row to zero?

Using logical indexing, you can use 'U>5' to indicate which rows/columns you want to change:

U = sum(A, 2); % sum of rows
disp(U > 5); % display 0-1 values where condition of false/true
A(:, U>5) = 0; % columns
A(U>5, :) = 0; % rows