Closed shipengfei92 closed 3 years ago
Hi @shipengfei92,
De_Boor's smoothing cubic spline for p=1 converges to the "natural" spline interpolant. It seems we cannot change boundary conditions in proposed approach and used solution method.
See https://en.wikipedia.org/wiki/Smoothing_spline#De_Boor's_approach
Maybe the solution can be found in the following document: http://monoceros.physics.muni.cz/~jancely/NM/Texty/Numerika/CubicSmoothingSpline.pdf
But I haven't looked in detail yet.
Thanks @espdev , I went through this paper, and it has the same idea with De_Boor's method, it will use the natural boundary to create a symmetric parameter matrix R, which can be used to simply the optimizing function, will try to generate a similar symmetric matrix based on the clamped constraints.
Thanks @shipengfei92 ,
If you find a solution for smoothing cubic spline with clamped boundary, PR welcome!
Sure @espdev , thanks for your efforts!
Hi! Thanks for this great package. Currently the smoothing spline is based on natural boundary, is there any possibility we can smooth a clamped spline?