Numerical Methods for Computer Science

Solve the least square problems using (QR, SVD, system of normal equations (solving the linear system with the Cholesky factorisation), Truncated SVD, LGK bidiagonalization (implemented for sparse matrices), gradient descent) and compare them. Shur decompositions of covariance or correlation matrices but you need to appropriately use the SVD factorization of the data matrix (or the standardized data matrix, respectively).