I have a very big matrix (hundreds of milions of entries) that holds a little number of values, and lots of zeros; that's why I thought to use a sparse matrix. There are plenty of libraries that save a sparse matrix (armadillo, openCV, eigen, sparselib++, so on and so forth), but I need to calculate its pseudo-inverse. Cited libraries have both methods, for computing sparse matrix and pseudo-inverse, but they didn't specify if they compute the pseudo-inverse OF a sparse matrix.
Could you please suggest me a library that computes a pseudo-inverse of sparse matrix? Sparselib++ of course will, but I wonder which is the most optimized way to achieve that task.
I'm using in my project C++ with openCV and PCL.
EDIT: I've looked on Eigen's faq, and it looks like they didn't implemented (yet?) the moore-penrose pseudo-inverse for sparse matrix. http://eigen.tuxfamily.org/index.php?title=FAQ#Is_there_a_method_to_compute_the_.28Moore-Penrose.29_pseudo_inverse_.3F