Solve nonlinear curve-fitting (data-fitting) problems

Hi all, I have a problem with Solve nonlinear curve-fitting (data-fitting) lsqcurvefit. In fact an error appears to me :

if true
  Error using lsqcurvefit (line 248)
Function value and YDATA sizes are not equal.
Error in glob_lv (line 305)
end

The code is :

if true
 function yy = func(m,xdata)
yy= [(m(1,1) * xdata(1,:) + m(1,2)* xdata(2,:) + m(1,3) * xdata(3,:) + m(1,4)) / (m(3,1) * xdata(1,:) + m(3,2)* xdata(2,:) + m(3,3) * xdata(3,:) + m(3,4)) ;
 (m(2,1) * xdata(1,:)+ m(2,2)* xdata(2,:) + m(2,3) * xdata(3,:) + m(2,4)) / (m(3,1) * xdata(1,:) + m(3,2)* xdata(2,:) + m(3,3) * xdata(3,:) + m(3,4))] ;
end
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
xda= [ x1 , x2 , x3 , x4 , x5 , x6 , x7 , x8 , x9 , x10 , x11 , x12]
yda= [ y1 , y2 , y3 , y4 , y5 , y6 , y7 , y8 , y9 , y10 , y11 , y12]
zda= [ z1 , z2 , z3 , z4 , z5 , z6 , z7 , z8 , z9 , z10 , z11 , z12]
xdata= [ xda ; yda  ;zda]
ydata = [ u1 ,u2 ,   u3 ,u4 ,  u5 ,u6 ,  u7 ,u8 ,  u9 ,u10 ,  u11 ,u12,
        v1 ,v2 ,   v3 ,v4 ,  v5 ,v6 ,  v7 ,v8 ,  v9 ,v10 ,  v11 ,v12 ]
m0=m_svd;
options = optimoptions('lsqcurvefit','Algorithm','levenberg-marquardt');
lb = [];
ub = [];
m = lsqcurvefit(@func,m0,xdata,ydata,lb,ub,options);
end

Any help ?

Answer 1

[Star Strider]

That error usually means that your data are in rows (or columns) and the output of your objective function are in columns (or rows), so they don’t match in dimension. Your objective function otherwise appears correct, so consider transposing the output matrix in it.

For example:

function yy = func(m,xdata)
yy= [(m(1,1) * xdata(1,:) + m(1,2)* xdata(2,:) + m(1,3) * xdata(3,:) + m(1,4)) / (m(3,1) * xdata(1,:) + m(3,2)* xdata(2,:) + m(3,3) * xdata(3,:) + m(3,4)) ;
 (m(2,1) * xdata(1,:)+ m(2,2)* xdata(2,:) + m(2,3) * xdata(3,:) + m(2,4)) / (m(3,1) * xdata(1,:) + m(3,2)* xdata(2,:) + m(3,3) * xdata(3,:) + m(3,4))].' ;
end

Note the addition of the (.') operator at the end:

... * xdata(3,:) + m(3,4))].' ;

so it takes the simple transpose. (The (') operator takes the complex-conjugate transpose, although the two are equivalent for real arrays.)

I cannot run your code to check it, so you will have to see if that solves the problem.