fitting data to implicit function

I have an experimental data with first column an independent variable and the second column a dependent variable. I need to estimate 2 parameters k and b. My model is actually a system o equations. The code of the model is below:

function y=calc_bg1(x,z) for ii=1:length(z) syms b gbg k n p r rgbg1 rgbg2 rgbg3 rgbg4 positive b=x(1); k=x(2); n=z(ii); [gbg,p,r,rgbg1,rgbg2,rgbg3,rgbg4]=solve(4*r*gbg == (k)*rgbg1, 3*rgbg1*gbg == 2*(k)*b*rgbg2, 2*rgbg2*gbg == 3*(k)*(b^2)*rgbg3, 1*rgbg3*gbg == 4*(k)*(b^3)*rgbg4, r+rgbg1+rgbg2+rgbg3+rgbg4 == 7*166e-9, gbg+rgbg1+2*rgbg2+3*rgbg3+4*rgbg4 == n*166e-9, (.97*rgbg4+0.03*7*166e-9)*52806306*0.19 == p*1e-3 ,gbg,p,r,rgbg1,rgbg2,rgbg3,rgbg4); pp(ii)=double(p); clear gbg n p r rgbg1 rgbg2 rgbg3 rgbg4 end y=pp'; When I am trying to do the fit with lsqcurvefit lsqcurvefit(@calc_bg1,[.4 7e-6],test1(:,1)',test1(:,2)) Improper assignment with rectangular empty matrix.

Error in calc_bg1 (line 9) pp(ii)=double(p);

Error in lsqcurvefit/objective (line 261) F = feval(funfcn_x_xdata{3},x,XDATA,varargin{:});

Error in snls (line 329) newfvec = feval(funfcn{3},xcurr,varargin{:});

Error in lsqncommon (line 155) [xC,FVAL,LAMBDA,JACOB,EXITFLAG,OUTPUT,msgData]= ...

Error in lsqcurvefit (line 253) [xCurrent,Resnorm,FVAL,EXITFLAG,OUTPUT,LAMBDA,JACOB] = ...

Can anybody help?

[Karan Gill]

Please format your code. It's really hard to read.

[Daniel Yakubovich]

is it better?

[Daniel Yakubovich]

the code of the function is:

 function y=calc_bg1(x,z)
for ii=1:length(z)
syms b gbg k n p r rgbg1 rgbg2 rgbg3 rgbg4 positive
b=x(1);
k=x(2);
n=z(ii);
[gbg,p,r,rgbg1,rgbg2,rgbg3,rgbg4]=solve(4*r*gbg == (k)*rgbg1, 3*rgbg1*gbg == 2*(k)*b*rgbg2, 2*rgbg2*gbg == 3*(k)*(b^2)*rgbg3, 1*rgbg3*gbg == 4*(k)*(b^3)*rgbg4, r+rgbg1+rgbg2+rgbg3+rgbg4 == 7.10838649776828*166.057788110262e-9, gbg+rgbg1+2*rgbg2+3*rgbg3+4*rgbg4 == n*166.057788110262e-9, (.97*rgbg4+0.03*7.10838649776828*166.057788110262e-9)*52806306.6954643*0.1960749100506533 == p*1e-3 ,gbg,p,r,rgbg1,rgbg2,rgbg3,rgbg4);
%pp(ii)=double(vpa(p))
pp(ii)=double(p);
clear gbg n p r rgbg1 rgbg2 rgbg3 rgbg4
end
y=pp';

and the fitting is done with: p=lsqcurvefit(@calc_bg1,[.4 7e-6],test1(:,1)',test1(:,2)) where test1 is a 2 column matrix