error when trying to fit a custom function using cftool

Hi, I'm trying to fit data to a custom exponential function a*exp(-b*x)+c. No fit is generated and I get the following error in the Results window: "Dimensions of matrices being concatenated are not consistent." I've tried generating clean, simple data to fit:

aa=2;
bb=0.5
cc=0.1;
xx=0:0.1:10;
yy=aa*exp(-bb*xx)+cc;

,cleaning all other variables (by opening a new instance of Matlab), and trying other custom functions, but nothing works...

It's worth mentioning that pre-defined fits (Polynomial, Exponential, etc.) can be generated.

I am using version R2012b.

[Elizabeth Jones]

I'm having the same problem now in R2014b. Did you ever find a solution?

Thanks! Elizabeth

Answer 1

[Timothy Wiggin]

I had the same problem, I'm going to document how I fixed it in case it comes up for anyone else. Skipping ahead, try restoredefaultpath. I'm using Matlab R2017a.

The root of the problem is parsing the custom equation. Running this code:

xx = 1:20;
yy = 0.5 * exp(2*xx);
fit(xx' , yy' , 'a * exp(b * x)' )

Produces this output:

Error using horzcat
Dimensions of matrices being concatenated are not consistent.

Error in symvar>findrun (line 163)
d = diff([0 x 0]);

Error in symvar (line 26)
[b,e] = findrun(s==' ');

Error in fittype>iCreateCustomFittype (line 463)
variableNames = symvar( obj.expr );

Error in fittype>iCreateFittype (line 351)
    obj = iCreateCustomFittype( obj, varargin{:} );

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

Error in fit>iFit (line 157)
    model = fittype( fittypeobj, 'numindep', size( xdatain, 2 ) );

Error in fit (line 108)
[fitobj, goodness, output, convmsg] = iFit( xdatain, ydatain, fittypeobj, ...

symvar crashes because the vector s is a column vector, so the s==' ' expression produces a column vector, and the concatenation expects a row vector. The problem is somehow related to the Matlab path, because running:

restoredefaultpath

Fixes the problem.

[Alex R.]

Yes and no. Check whether you have a function called "vectorize" (which probably outputs a column vector) in one of your path included folders. Matlab will use that instead of the original "vectorize" function (which is not the most suitable name name if I was honest for what it's supposed to do).

I know this question is old, but I just commited that same mistake ...

Answer 2

[Salaheddin Hosseinzadeh]

Hi Guys,

Well done for providing the equations.

It worked fine for me, I attached an image

Or