How to solve an error while using the command 'svmtrain'?

Hi,

I am trying to classify a set of numeric data into two classes using SVM. There are 8 rows(observations) and 6 columns(features) in the training data. The target is a numeric column matrix of size 8x1.

The training data is

data_train =

    0.0713   -0.0810    0.1523    0.0141    0.0135   13.1808
    0.0789   -0.0974    0.1762    0.0161    0.0156   10.2324
    0.0752   -0.0842    0.1595    0.0156    0.0151   10.9373
    0.0905   -0.0940    0.1845    0.0184    0.0180   13.1222
    0.8479   -0.7886    1.6365    0.0721    0.0720   27.8630
    0.5273   -0.8057    1.3330    0.0750    0.0749   28.7584
    0.4353   -0.5557    0.9911    0.0596    0.0595   25.1581
    0.4944   -0.5167    1.0110    0.0524    0.0523   22.8835

The target data is

groups =

     1
     1
     1
     1
     2
     2
     2
     2

I use the following code

cp = classperf(groups);

svmStruct = svmtrain(data_train, groups,'Kernel_Function', 'rbf');

I get the following error message:

??? Undefined variable "internal" or class "internal.stats.getargs".

Error in ==> svmtrain at 294 [eid,errmsg,kfun,optimMethod, plotflag, polyOrder, ... mlpParams, boxC, rbf_sigma, autoScale, opts,... tolkkt, kktvl,kerCL,... kfunargs qpOptsInput, smoOptsInput] = ... internal.stats.getargs(pnames, dflts, varargin{:});

Could you please let me know what does this error mean, and how to overcome it?

I use MATLAB R2011a 64 bit version with Windows 7.

Thanks in advance.

Answer 1

[Shashank Prasanna]

This sounds like a path issue. Can you try the following:

>> which -all internal.stats.getargs

You should see an output that tells you where it is. If you seen an output that says not found, you might have to reset your path.

>> restoredefaultpath
>> rehash toolboxcache

[Sutapa]

Thanks, it worked.