Simple SVM classification code

Hi,

I have Matlab2013a, and use the following code for classifying my data (it is mostly taken from http://stackoverflow.com/questions/3070789/example-of-10-fold-svm-classification-in-matlab). My feature matrix is 84x19900 (84 subjects, each subjects having a row vector with 19 900 features), and class is a 84x1 logical array (1-42 are 1s, 42-84 are 0s). Yet my code throws an error (see end):

load feature load class

k = 7; %number of folds cvFolds = crossvalind('Kfold', class, k); cp = classperf(class);

for i = 1:k                                  %# for each fold
    testIdx = (cvFolds == i);                %# get indices of test instances
    trainIdx = ~testIdx;                     %# get indices training instances

      %# train an SVM model over training instances
      svmModel = svmtrain(feature(trainIdx,:), class(trainIdx), ...
                   'Autoscale',true, 'Showplot',false, 'Method','QP', ...
                   'BoxConstraint',2e-1, 'Kernel_Function','rbf', 'RBF_Sigma',1);

      %# test using test instances
      pred = svmclassify(svmModel, feature(testIdx,:), 'Showplot',false);

      %# evaluate and update performance object
      cp = classperf(cp, pred, testIdx);
  end

%# get accuracy
cp.CorrectRate

%# get confusion matrix
%# columns:actual, rows:predicted, last-row: unclassified instances
cp.CountingMatrix

Error:

Error using svmclassify (line 53)
The first input should be a struct generated by SVMTRAIN.

Error in classifyAutismTD (line 18)
    pred = svmclassify(svmModel, feature(testIdx,:), 'Showplot',false);

What is wrong?

[Hamoon]

Check this code first:

 feature = rand(84,19900);

 class = rand(1,84)>.5;
 k = 7; %number of folds 
 cvFolds = crossvalind('Kfold', class, k); 
 cp = classperf(class);

 for i = 1:k                                  %# for each fold
     testIdx = (cvFolds == i);              %# get indices of test instances
     trainIdx = ~testIdx;                  %# get indices training instances
     %# train an SVM model over training instances
     svmModel = svmtrain(feature(trainIdx,:), class(trainIdx), ...
         'Autoscale',true, 'Showplot',false, 'Method','QP', ...
         'BoxConstraint',2e-1, 'Kernel_Function','rbf', 'RBF_Sigma',1);
     %# test using test instances
     pred = svmclassify(svmModel, feature(testIdx,:), 'Showplot',false);
     %# evaluate and update performance object
     cp = classperf(cp, pred, testIdx);
 end
 %# get accuracy
 cp.CorrectRate
 %# get confusion matrix
 %# columns:actual, rows:predicted, last-row: unclassified instances
 cp.CountingMatrix

Are you getting the same error here? if the answer is yes, look to the variable svmModel in your workspace, what's its value?

[MiauMiau]

Many thanks! So, yes, I still got the error. And the problem was the following: I have libsvm in the same folder, and it looks like some functions are named the same, hence the error. Now I removed the folder from the path, and everything is fine..many thanks nevertheless!

[the cyclist]

You might consider copying this last comment of yours into an answer, and then accepting it. This might help other people in the future with a similar problem.

[Vania krm]

Dear @MiauMiau

I have also this error and I can not solve it. Can you explain more how I can check the libsvm folder?