'Having a set of 3d curved plane peaces to learn, how to get a new curve having some parts of its data?

So I have a set of 3d curved planes defined by a set of points [[x, y, z]] each, all near zero, yet some with extremely similar yet different shape. I can pretrain/train/fit/somethin on this data for long time.

At runtime I get two new sets:

  • a context - set of known [[x, y, z]] points of a new given plane
  • set of [[x, y]] points for which I want to estimate closest [[z]] with respect to my original curve patterns and given context

What would be the algorithm to fit to such problem that would try to fit a new curved plane to patterns from the past looking at given data known about the new curved plane?

algorithmmachine-learningscikit-learncurve-fitting


Sources

This article follows the attribution requirements of Stack Overflow and is licensed under CC BY-SA 3.0.

Source: Stack Overflow

Solution Source

Related Questions