Add regularization for BSpline #348
Conversation
src/b-splines/prefiltering.jl
Outdated
| sz = size(ret) | ||
| first = true | ||
| if ndims(ret) > 1 | ||
| @warn "Smooth BSpline only available for Vectors, fallback to non-smooth" |
There was a problem hiding this comment.
How about restricting interpolate to AbstractVector when λ and k are supplied?
There was a problem hiding this comment.
It should work for higher dimensions too, it is just not implemented yet.
When implemented, it will be easier to change then.
src/b-splines/prefiltering.jl
Outdated
| M, b = prefiltering_system(TWeights, eltype(TCoefs), sz[1], degree(it)) | ||
| ### TEST REGULARIZATION | ||
| n = sz[1] | ||
| Q = Matrix(diffop(TWeights, n, k)) |
There was a problem hiding this comment.
Do you have to enforce dense here? I'm wondering what happens on a large array.
There was a problem hiding this comment.
I had to transform to Dense matrix because I was getting an error running Cholesky with sparse matrices (see next line). I don't know enough about this type of optimization to solve it. But yes, there is no reason for it to be dense.
|
I added tests, but they are maybe too minimal. I have no idea what is important to check. If somebody can comment on it that would be great. |
|
This looks good to merge this now. Could we update the base branch so this can automatically merged? |
|
It would be good to try optimizing the code before (but I don't know enough to do it), like for instance @timholy 's comment about transforming to a dense Matrix, I think it could be avoided. |
|
Tests are passing here and this is mergeable, so I'm merging for v0.13.1 prospectively. Feel free to submit another PR for further optimization slated for v0.13.2 or earlier. |
Closes #254
Implements a regularized BSpline.
To implement a regularized spline on unequal-grid, the x-axis should be provided from the beginning, it cannot work with defining a ScaledSpline afterwards.
No tests for now and minimal documentation.