compute Eigenvalues/EigenVector

I am looking for code that computes the Eigenvalues and Eigenvectors using the Accelerate Sparse Matrix library.

There's a very similar question on SO. If your matrix is symmetric, then following approach will work:

  1. SparseFactor the matrix with the LDL^T factorization
  2. Use SparseCreateSubfactor to get an SparseOpaqueSubfactor objects representing D and L
  3. SparseMultiply those objects by the identity matrix to get D and L and as dense matrices. You can read off your eigenvalue/vector pairs from that.

If the matrix isn't symmetric, then...there might be a way to use the QR factorization to indirectly say something about the eigenvalues. But the QR factorization doesn't tell you the eigenvalues straight-up.

compute Eigenvalues/EigenVector
 
 
Q