I'm trying to solve a linear system of equations in Python with a very sparse, relatively big matrix (size around 10 x 10000). Using Numpy, it was easy to get a numerical solution using the least squares method. However, it has next to no zeros in it, meaning the result is a linear combination of thousands of vectors.
I also tried using SciPy, so that I could try to tinker with the parameters to maximize the number of zeros, but I don't know it well at all, and am not sure if there's a way to do it with so many variables at once.
Basically, my question then is: is there a way to either (1) specify to Numpy that I want as many zeros as possible in the solution (instead of the usual norm minimization); or (2) solve big systems directly in SciPy?
Thanks in advance for any help!