How to do this: Multivariate Interpolation with random data (non-gridded)

Dear all,

I have a question needing your help. Thanks in advance. I have a random data (not grids), (x1, x2, y), hypothetically generated from some unknown relationship y=f(x1,x2). I want to do interpolation and evaluate the value at some known points (x10, x20). Is there any way to do it?

Or any suggestion for fine approximation based on (x1,x2,y) to predict f(x10,x20) with reasonable accuracy? I have tried polynomial approximation (non-piecewise) and the result was not good enought. Thanks a lot.

Best, Zppine

Answer 1

[Andrei Bobrov]

Please use scatteredInterpolant

Answer 2

[George Papazafeiropoulos]

vq = griddata(x1,x2,y,x10,x20)

hth,

George Papazafeiropoulos