Recommended routines for interpolation in `scipy.interpolate`

recommendation_table.md

	kind	routne	continuity	comment	alias
	linear	numpy.interp	broken line	comes from numpy
	cubic spline	CubicSpline	2nd derivative
1D	monotone cubic spline	PchipInterpolator	1st derivative	non-overshooting
	non-cubic spline	make_interp_spline	(k-1)th derivative
	nearest	interp1d	none	kind='nearest',
				'previous', 'next'
-----------------
N-D curve	nearest, linear, spline	make_interp_spline	(k-1)th derivative	use N-dim y array
-----------------
	nearest
N-D regular grid	linear	RegularGridInterpolator		rectilinear grids	interpn
	splines		2nd derivatives
	monotone splines		1st derivatives	method='pchip'
-----------------
	nearest	NearestNDInterpolator
N-D scattered	linear	LinearNDInterpolator			griddata
	cubic (2D only)	CloughTocher2DInterpolator	1st derivatives

Nice list 👍 Maybe also add RBFInterpolator in N-D scattered?

@tupui yes, great idea! In the scipy.interpolate tutorial, which is where the canonic version of this table resides now :-).

ah yes cool. I missed/forgot that it went in.

Indeed, this was a wall of text PR, easy to miss parts of! BTW,the whole rbf section of the tutorial needs porting from rbf to rbfinterpolator. Would be even better if an rbf/rbfinterpolator user does it. Are you?

Truth to be told at the moment I am really only a user of stats.qmc 😅 But agreed we should at least update the code to use RBFInterpolator. I can do that quickly if you want.

Would be great actually!

Re stats.qmc --- this is a great piece of work, much appreciated!

Thanks 😃

And just opened scipy/scipy#17371