curve fitting with more then two independent variables

Hi I had a question about curve fitting,my Costume equation is f(x)=(a+b*x*z)/(1+c*x),the tricky point here is that z is not unknown it is a vector same size as x and just a b c are unknown constants,how is the solution?

Answer 1

[Miroslav Balda]

If the regression function is in right MATLAB notation

      f(x) = (a+b*x.*z)./(1+c*x) % see dots between x and z and before /!,

the coefficients a, b and c may be found by the application of the function LMFnlsq from File Exchange, see

      www.mathworks.com/matlabcentral/fileexchange/17534

for a nonlinear regression. The problem may be tranformed into linear one in unknown coefficients:

      f(x).*(1+c*x) = a+b*x.*z

The solution of this problem by a linear regression, a0, b0, c0 can serve as a good initial guess for the nonlinear regression of the original nonlinear problem.

[kesh shahi]

>> lookfor LMFnlsq

LMFnlsq not found.

>>

?I think it was omitted in new version of MATLAB(2012Ra)?