Problem with Fminsearch function

 % Load the data 

idxMiddle = blk.audDiff == 0;
audMiddle = [blk.visDiff(idxMiddle,1) blk.responseMade(idxMiddle)];
audMiddle = audMiddle(~audMiddle(:,2)==0,:); 
ccSym = blk.visValues;  
tn = arrayfun(@(x) length(audMiddle(audMiddle(:,1)==x,2)==2), ccSym); 
meanProb = arrayfun(@(x) mean(audMiddle(audMiddle(:,1)==x,2)==2) , ccSym);
reactionLength = blk.responseTime;

% % Do the fitting 
parstart = [3 0.5 0.5 0 0.1]; %Initial parameters
nfits = 10; %Number of iterative fits
minErrors = zeros(nfits,1); %Vector to keep track of errors with each fit
pars = cell(nfits,1); %Cell array to keep track of parameters for each fit

%Loop to fit data 10 times
for i=1:nfits
    erHandle = @(pars)(errorFun(pars ,[ccSym'; tn'; meanProb'; reactionLength])); 
      options = optimset('MaxFunEvals', 1000);
      pars{i} = fminsearch(erHandle, parstart, options);
      minErrors(i) = errorFun(pars{i}, [ccSym'; tn'; meanProb'; reactionLength]);
end
[L,iBestFit] = min(minErrors);
bestPars = pars{iBestFit};

I have a problem with this program, and I'm receiving these errors when I try to run it:

>> model
Error using vertcat
Dimensions of arrays being concatenated are not consistent.

Error in model>@(pars)(errorFun(pars,[ccSym';tn';meanProb';reactionLength]))

Error in fminsearch (line 200)
fv(:,1) = funfcn(x,varargin{:});

Error in model (line 52)
    pars{i} = fminsearch(erHandle, parstart, options);

What would you suggest me to do? Thank you in advance.

[Star Strider]

Use the size function to determine the dimensions of: ccSym, tn, meanProb, and reactionLength.

You can also use the whos function, for example:

x = rand(3,2);
whos('x')

  Name      Size            Bytes  Class     Attributes
  x         3x2                48  double   

that gives a bit more information.

Posting the results to a Comment here would help us to determine the problem.

[jasmine]

Hi,

Sorry for the late reply, I had to change some parts of the code as reactionLength was not the same size as the other ones.

This is what I did:

audMiddle_time = [blk.timeToWheelMove(idxMiddle) blk.responseMade(idxMiddle) blk.visDiff(idxMiddle,1)];audMiddle_time = audMiddle_time(audMiddle_time(:,2)==2,:); reactionLength = arrayfun(@(x) median(audMiddle_time(audMiddle_time(:,1)==x,2)==2), ccSym);

But I don't understand why do I get the reactionLength in a logical form. How can I change that? I still have the same errors, and I think it's because of this.

>> whos('tn')

Name Size Bytes Class Attributes

tn 9x1 72 double

>> whos('reactionLength')

Name Size Bytes Class Attributes

reactionLength 9x1 9 logical

>> whos('meanProb')

Name Size Bytes Class Attributes

meanProb 9x1 72 double

>> whos('ccSym')

Name Size Bytes Class Attributes

ccSym 9x1 72 double

[Torsten]

You forgot to transpose "reactionLength" in

[ccSym'; tn'; meanProb'; reactionLength]

[jasmine]

Oh, yeah, but I still have the same errors :(

[Torsten]

What about

erHandle = @(pars)(errorFun(pars ,[ccSym.'; tn.'; meanProb.'; (double(reactionLength)).']));