Is there a way to curve fit a nonlinear implicit function?

[Dr_Laine]

Hi

I would like to fit a least squares curve to a set of data points.  I see I can do this using general LSQ fit.vi or Levenberg-Marquadt, but these VIs seem to only handle implicit functions of the form y=f(x,A). 

What if I have a data set from a known implicit function, for example a circle:

(x-x0)^2 + (y-y0)^2 - r^2 = 0

How can I calculate the x0, y0 and r that minimizes the square error?  Can this be done using the L-M VI or a VI from a different toolkit?

Thank you.

[chilly_charly]

This can be considered as an optimization problem. For a circle, you could use the fitting on a sphere vi (in the math > optimization sub-palette), giving 0 as z coordinates.  See the attached vi.

[Dr_Laine]

Fantastic, worked like a charm!

Thanks.