Fit, fittype, fitoptions

Hello,

I have a two input, one output custom equation that I am trying to define the coefficients using field data. The code I am using is as below

fo = fitoptions('Method','NonlinearLeastSquares', 'Lower',[0 0 0 0 ]', 'Upper', [5 5 4000 200 ], 'StartPoint',[0.1 0.1 0 0], 'MaxIter',4000);
ft = fittype( @(a,b,c,d,m,n) a.*(exp(b.*m)-1)+c+d.*n.^2,'independent', {'m', 'n'},  'options', fo);
f = fit(  M, N,Q, ft);

M and N are variables. a, b,c,d need to be defined.

When I run it, I get the following error:

Error using fit>iParseOptionalArgs (line 965)
The properties, PROP1, PROP2, ..., in property-value pairs must be
strings.

Error in fit (line 105)
[useroptions, useroptargs, probparams] = iParseOptionalArgs(
varargin{:} );

Error in FirstTryWithFlowRate (line 106)
    f = fit(  M, N,Q, ft, fo);

Any ideas on what might be causing this? I do not have any problem when I have only one independent variable. I also don't have any problem when I do it through the Curve Fitting application. But I need to automate the process, so I need to do it via script.

Thanks

Answer 1

[Mete Mutlu]

After much deliberation, I found my mistake. The inputs M and N should be in bracket. So here would be the correct format.

f = fit([M, N],Q, ft);

The rest is the same.

For the record, if you have a more complicated fit or options, try setting it up in the curve fitting toolbox, then, File> Generate Code will generate the code for you that you can use later on.