[SciPy-Dev] Mass 1D interpolation, advice

[josef.pktd at gmail.com]

On Tue, Apr 14, 2015 at 6:39 PM, Matthew Brett <matthew.brett at gmail.com> wrote:
> Hi,
>
> On Tue, Apr 14, 2015 at 3:24 PM, Evgeni Burovski
> <evgeny.burovskiy at gmail.com> wrote:
>> What's the interpolation order?
>
> Cubic spline, linear, nearest neighbor in that order of frequency.
>
>> * Warren recently added nd functionality to splev:
>> https://github.com/scipy/scipy/pull/4059
>
> A quick scan suggested this allows an N-D array of x input values,
> whereas I want to evaluate a 1D array of x input values for 2- or N-D
> array of y values - is that right?

If you have the same 200 x values for all the y interpolation, then
you can reuse the projection matrix which is only 200x200. Similar
idea as in Savitsky-Golay.

It would be just one big dot product 200x200 dot 200x4000, after
creating the projection matrix

(not a how to use scipy answer)

Josef

>
>> * Have you tried polynomial interpolators? PPoly,
>> BPoly.from_derivatives might be what you are after
>
> Forgive my terrible ignorance, but is it easy to do cubic spline
> interpolation using these classes?
>
> Thanks for the reply,
>
> Matthew
> _______________________________________________
> SciPy-Dev mailing list
> SciPy-Dev at scipy.org
> http://mail.scipy.org/mailman/listinfo/scipy-dev