System of Equations with LSQCURVEFIT

I would like to fit my data to a system of equations. I keep getting an error of "LSQCURVEFIT requires all values returned by user functions to be of data type double."

Any suggestions in getting my code to run correctly?

kappa1 = @(x, xdata)x(1)*xdata*x(2); % Equation 1
kappa2 = @(x, xdata)2*sqrt(2)*x(1).^2*x(2)*(-2*x(3).*(xdata+(xdata+x(3)).*log10(x(3)./...
    (xdata+x(3))))); % Equation 2

E1 = char(kappa1) % Equation listing turn to character string; then combine
E2 = char(kappa2)
ES = (['{' E1 '; ' E2 '}'])
ESF = inline(ES)

k_2 = fcum(1:2,1:2); % ydata 2 by 2 matrix, two answers for each equation for two xdata points
T = AllDataTime; % two data time points
x0 = [.618 .088 .2]; % Initial guess values

[x, resnorm] = lsqcurvefit(ESF, x0, T, k_2)

Answer 1

[Walter Roberson]

lsqcurvefit cannot be used to fit to a system of equations. The output of the objective function must be the same size as the ydata, and cannot be a cell array.

The original poster also had difficulty with handling symbolic functions, but there is no point in posting a solution to that part as the overall goal cannot be met using lsqcurvefit

Answer 2

[Jing_KU]

Have you solved this problem?

Regards, Jing