what does fitoptions 'Method' = 'None' mean?

I just looked up the defaults for fitoptions. For 'Method' the default appears to be 'None'.

>> fitoptions

ans =

    Normalize: 'off'

      Exclude: [1x0 double]

      Weights: [1x0 double]

       Method: 'None'

But what does that mean? It has to use SOMETHING to determine whether it has found the best fit. I can't seem to find out what, though. I'm very curious about this now.

Answer 1

[David Rosin]

Alright! This is it:

        Normalize: 'off'
          Exclude: []
          Weights: []
           Method: 'NonlinearLeastSquares'
           Robust: 'Off'
       StartPoint: [1×0 double]
            Lower: [1×0 double]
            Upper: [1×0 double]
        Algorithm: 'Trust-Region'
    DiffMinChange: 1.0000e-08
    DiffMaxChange: 0.1000
          Display: 'Notify'
      MaxFunEvals: 600
          MaxIter: 400
           TolFun: 1.0000e-06
             TolX: 1.0000e-06

So as it turns out, it's a nonlinear least-squares approach with the trust region algorithm. The options-object seems to be always initialized like this. After that, the user-input is checked for an options-arguments to overwrite/edit it with.

Thanks again for pointing me in the right direction!

Answer 2

[the cyclist]

I expected the documentation to be a little more explicit about how the fitting functions handle things when Method is 'None', but it doesn't seem to be. I wonder if different functions have different default behavior.

I don't have the Curve Fitting Toolbox, so I can't explore. What I would do next would be to edit the actual function you are using to fit, search on the term 'None', and see if you can glean what it does with that. (Of course, be careful to not modify the function when in edit mode.)

[David Rosin]

Thank you for your reply.

That does sound like a good way to go about it. I'll look into that.