least square fitting of multiple variable equation (error: too many input arguments)

7 views (last 30 days)
I have a data file constituting 2 lines (EXP_x, EXP_y) and want to fit it with the equation mentioned in the second paragraph which has two variables (w, a).
here is the code
EXP=textread('filename')
global EXP_x
global EXP_y
EXP_x=EXP(:,1);
EXP_y=EXP(:,2);
w0=[0 0.5];
lb=[0 0.5];
ub=[20 2];
x=lsqcurvefit(@distribution,w0,EXP_x,EXP_y,lb,ub);
function calculation=distribution(w,a)
global EXP_x
global EXP_y
gamma=@(t) integral(t.^(w-1).*exp(-t),0,inf);
calculation=(a/((w^(3/2))*gamma)*exp(-(EXP_x/w).^a))
end
it doesn't work. Error: "lsqcurvefit too many input arguments in that case"
If it is not the proper case to use the lsqcurvefit, please tell me the replacement method.
Thank you.

Accepted Answer

Torsten
Torsten on 6 Mar 2022
Edited: Torsten on 6 Mar 2022
x=lsqcurvefit(@(w)distribution(w(1),w(2)),w0,EXP_x,EXP_y,lb,ub);
instead of
x=lsqcurvefit(@distribution,w0,EXP_x,EXP_y,lb,ub);
and
function calculation=distribution(w,a)
global EXP_x
global EXP_y
gammah = gamma(w);
calculation=(a/((w^(3/2))*gammah)*exp(-(EXP_x/w).^a))
end
instead of
function calculation=distribution(w,a)
global EXP_x
global EXP_y
gamma=@(t) integral(t.^(w-1).*exp(-t),0,inf);
calculation=(a/((w^(3/2))*gamma)*exp(-(EXP_x/w).^a))
end

More Answers (0)

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!