lsqcurvefit help: Error using asset. FUN must have two input arguments.

I'm trying to work through a curve fitting problem and I haven't been able to resolve the following coding error. Unfortunately, I haven't done much curve fitting as of yet, so any help is greatly appreciated!

PV = readtable('clean_data.csv');
NV = readtable('noisy_data.csv');

% PV is clean
lambda_p = table2array(PV(:,1));
I_p = table2array(PV(:,2));

% NV is noisy
lambda_n = table2array(NV(:,1));
k_n = lambda_n./(2*pi);
I_n = table2array(NV(:,2));

w_e = 2.8e-8;

% Curve fitting with v_hat in Fourier space

v_hat_p = abs(fft(I_p));
k_p = lambda_p./(2*pi);

v_hat = @(A_v,w_g,w_l) A_v.*exp(-(w_g.^2*k_p.^2)/(16*log(2)) - (w_l.*abs(k_p))./2);
A_v = 100; w_g = 0.01; w_l = 0.01; % Initial guesses

x0 = [A_v, w_g, w_l]; % initial guesses
[x_fit,y_fit] = lsqcurvefit(v_hat,x0,k_p,v_hat_p);

Error using assert

FUN must have two input arguments.

Error in lsqcurvefit (line 229)

assert(numberInput == 2, message('optimlib:lsqcurvefit:NotEnoughInputArg'));

Error in myfilename (line 29)

[x_fit,y_fit] = lsqcurvefit(v_hat,x0,k_p,v_hat_p);

Answer 1

[Torsten]

Replace the line

v_hat = @(A_v,w_g,w_l) A_v.*exp(-(w_g.^2*k_p.^2)/(16*log(2)) - (w_l.*abs(k_p))./2);

by

V_hat = @(A_v,w_g,w_l,k_p) A_v.*exp(-(w_g.^2*k_p.^2)/(16*log(2)) - (w_l.*abs(k_p))./2);
v_hat = @(x,xdata) V_hat(x(1),x(2),x(3),xdata)

And the division by 16*log(2) is superfluous - this can be compensated by fitting w_g accordingly.

[Richard Schure]

Thanks so much! I did try to add the k_p variable to the original and that didn't work, but now I see how separating the equations fixes the problem.