Im looking to create a matrix of rank k.
The dimension of the matrix is m x n. The input k satisfies that condition that k < min(
Well, a trivial method is to produce a matrix that looks like:
1 0 0 0 0
0 1 0 0 0
0 0 1 1 1
0 0 0 0 0
i.e. k columns of the identity matrix, then repeat the last column n-k times (or m-k times, depending on orientation).
It's not really so clear what you are aiming for.
But in order to create a matrix B with specific rank k, from a matrix A (with rank at least k), you may like to utilize svd and proceed like:
>>> A= rand(7, 5);
>>> rank(A)
ans = 5
>>> [U, S, V]= svd(A);
>>> k= 3;
>>> B= U(:, 1: k)* S(1: k, 1: k)* V(:, 1: k)';
>>> rank(B)
ans = 3
A matrix of rank 1 can be created by the outer product of two vectors, for example:
A = randn(10,1) * randn(1,10);
Add together k of these and you will have a matrix of rank k. Like this:
>> A = zeros(10);
>> for i = 1:4, A = A + randn(10,1) * randn(1,10); end
>> rank(A)
ans = 4