How can I implement lsqcurvefit function on my equations?

Question

Ceyhan TURKMEN 2020-10-12

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/611346-how-can-i-implement-lsqcurvefit-function-on-my-equations

评论： Jon 2020-10-12

Hello, I have a nonlinear equation system of 5 variables (complex-valued) and I want to try “lsqcurvefit” function. I want my output to be as close to zero as possible so, should "ydata" must be equal to my expected values or a row vector of 5 zeros? Sorry if it's a dumb question but I am confused. Thanks for the help.

objfcn = @(x,xdata) [...Equations...];
  
x0 = Initial_Values(1,:)';
xdata = Measurements(1,:)';
ydata = Exp_XValues(1,:)';
[x,fval] = lsqcurvefit(objfcn, x0, xdata, ydata, [], []);

2 个评论
显示无隐藏无

Matt J 2020-10-12

You mean you want the x(i) to be as close to zero as possible? You think the optimal parameters are non-unique?

Ceyhan TURKMEN 2020-10-12

Actually, I search optimum x(i) values, whose roots of the equation set.

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Jon 2020-10-12

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/611346-how-can-i-implement-lsqcurvefit-function-on-my-equations#answer_511426

编辑：Jon 2020-10-12

From your description, it sound like you want to find the solution to a system of 5 equations (in I assume 5 unknowns). In this case I think fsolve would be more appropriate. lsqcurvefit is for the situation where you are trying to find the coefficients of a nonlinear expression to give the best fit to an observed data set. In particular this would be the situation where you have typically many more equations than unknowns, and your unknowns are the parameters of some nonlinear equation. So for example if you were trying to fit an equation y = a*x^b + c and you wanted to find the best values for a,b and c.