Why doesn't fittype, in the curve fitting toolbox, see independent variable in input list of called function?

I want to use a fit model in the curve fitting toolbox complicated enough to warrant an m-file, but can't get fittype to recognize the independent variable. The same error is generated using the simple function given in the Help example:

>> g = fittype( @(a, b, c, x) a*x.^2+b*x+c )

g = 

       General model:
       g(a,b,c,x) = a*x.^2+b*x+c

Suppose I name the model function:

>> f=@(a,b,c,x) a*x.^2+b*x+c

f = 

      @(a,b,c,x)a*x.^2+b*x+c

It evaluates nicely:

>> f(1,1,1,1)

ans =

       3

but throws an error in fittype:

>> fittype( @(a, b, c, x) f)

Error using fittype>iAssertIsMember (line 1082)

*Independent variable x does not appear in the equation expression.*

Error in fittype>iTestCustomModelParameters (line 728)

iAssertIsMember( obj.indep, variables, 'curvefit:fittype:noIndependentVar' );

Error in fittype>iCreateFittype (line 370)
    iTestCustomModelParameters( obj );

Error in fittype (line 328)
                obj = iCreateFittype( obj, varargin{:} );

Naming the independent variable:

>> fittype( @(a, b, c, x) f,'independent','x')

gives the same error.

An m-file also gives the same error. I delete f, save f.m containing:

function ff = f(a,b,c,x)

ff = a*x.^2+b*x+c;

and execute:

>> fittype( @(a, b, c, x) f,'independent','x')

and get the same error as above.

Answer 1

[Steven Lord]

See my response to your posting of this question in the newsgroup.

[RAJEEV]

Dear Steve,

I am having same problem can you please elaborate the cause. The link you provided is no longer working.

Answer 2

[Larry]

fittype( @(a, b, c, x) f *(a,b,c,x)*)

works w/ f.m.