I am trying to use the NationalInstruments.Analysis.Enterprise assembly to do a fit of a custom, nonlinear function with two independent variables and five parameters. The function is:
Z = A + 1 / (1 / (X / 256 * B + 3 * C) + 1 / (Y / 1024 * D + 2 * E))
Where Z is the output, X and Y are the independent variables, and A, B, C, D, and E are the parameters I want to estimate. I also have best guess values for these parameters to minimize the iterations necessary for the least squares to converge if that is an option.
I see there is a CurveFit class but it looks like it only supports a single independent variable X with an output Y. Is there another class I'm not seeing that would support two independent variables? Or is there a way to use the least squares directly by creating my own cost function? Any help is greatly appreciated.
