How can I use parameters from workspace in custom equation using fittype function?

I'd like to write the coefficient, in the custom equation, in literal form, keeping A as the solution of the curve fitting

function A = coffin_coefficient(X,Y)

[xData, yData] = prepareCurveData( X, Y );

ft = fittype( 'A*x^(-6.9397)*exp(0.8084/((8.6173e-05)*(120+273.15)))', 'independent', 'x', 'dependent', 'y' );
opts = fitoptions( 'Method', 'NonlinearLeastSquares' );
opts.Algorithm = 'Levenberg-Marquardt';
opts.Display = 'Off';
opts.StartPoint = 0.1;

Answer 1

[Guillaume Le Goc]

If you want to fit y(x) to find the parameter A, now you have set the fittype and fitoptions for the fit, you just need to perform the fit :

f = fit(xData, yData, ft, opts);

And for your function to return the found A, you can access it with :

A = f.A;

You might want to get the 95% confidence interval for the value :

ci = confint(f);

[Carmelo Barbagallo]

Guillaume thank you for your reply, but I have already performed the curve fitting. My problem is to replace the numbers in the equation inside "fittype" with parameters present in the workspace.

[Carmelo Barbagallo]

!!! I DID IT !!!

I have used "eval" function

function A = coffin_coefficient(deltaT,Nf,n,Ea,kb,Tm)

[xData, yData] = prepareCurveData( deltaT, Nf );

eval(['ft = fittype( ''A*x^(-',num2str(n),')*'...
    'exp(',num2str(Ea),'/((',num2str(kb),')*(',num2str(Tm),'+273.15)))'','...
    ' ''independent'', ''x'', ''dependent'', ''y'' );'])

opts = fitoptions( 'Method', 'NonlinearLeastSquares' );
opts.Algorithm = 'Levenberg-Marquardt';
opts.Display = 'Off';
opts.StartPoint = 0.1;

Probably there are better solution, but it seems to work.

However I am interested in a simpler solution

[Guillaume Le Goc]

Sorry, I did not understand what your problem was.

I generally read that the eval function is generally not recommended, especially here where it is of no use.

Since your variables n, Ea... are defined in the workspace from which you call your coffin_coefficient function, you can simply pass them as input arguments in your function, as you did. Now, those variables exist within the function, and you can simply define your fitting function normally as you did in the first place, replacing numerical values by their variable names.

In fact,

eval(['ft=a+b+c;']);

does the same thing than

ft=a+b+c;

but is way slower.

In short, now that you pass all the variables to your function, you can use them as you would in a script/command window, ie. using the workspace you have access to.