(Matrix R^m x d
s = subset(1:m,d))
and correspend this math notation.
Is there any funtions for these
This is too abstract for me. Could you please translate at least something of it to Julia?
Do you want a basis (i.e. generating set) for the right nullspace (i.e. right kernel) of a matrix over the reals? You can use the nullspace function for this.
3 Likes
The basis is not unique, of course.
Julia is giving an orthonormal basis from the SVD. The Python right_kernel function you are calling (from sage?) is giving a non-orthonormal basis that looks like it comes from a reduced row-echelon form algorithm, which is generally not used in numerical calculations (as opposed to symbolic math or hand calculations) because it is not robust to roundoff errors.
1 Like