Matlab curve fitting coefficient

0
votes

I have a question about curve fitting coefficient in matlab,

how can I get the exact value stored in p1, p2, p3, etc, as seen in the Matlab Curve Fitting Toolbox Interface:

curve fitting interface

I'm currently working on making the f(x) function from some known data and using a big exponential number type produced a lot of error to the function because it means the number rounded heavily.

1
matlabcurve-fitting
You have a warning about the equation being badly conditioned, you might want to tick the "center and scale" option... – Wolfie
I actually didn't know what that mean... I kind of new to Matlab and this community... I just started it like three days ago to do some graph (picture) to function converting... I currently using image processing to convert the image to a binary matrix so I can extract the graph line from the matrix and using some line of code to convert it's location in matrix to the actual x and y data from the graph, and the last part is using curve fitting to the data tp find the function of that graph... sorry if i you cant understand my explanation, english is not my native language – Sayid Ibrahim
but if you understand, do you have any other method to achieve that same purpose? it doesn't have to be matlab, thank you – Sayid Ibrahim

1 Answers

1
votes

You can save your fit to the workspace from the "Fit" menu

menu

With the default export options, you'll get a variable called fittedmodel which has the properties p1, p2, p3, ..., which are the values of your coefficients

>> x = linspace(0,10,100);
>> y = rand(size(x))*5 + x.^2;
>> % Use curve fitting app and export 'fittedmodel'
>> fittedmodel.p1
   1.034293219979936
>> fittedmodel.p2
  -0.308627984704966
>> fittedmodel.p3
   2.786188360368544