How to do cross validation SVM classifier

Question

I want to perform a decoding by applying an SVM classifier to a data matirx S, the size of which is 1089*43093,and the prediction accuracy of the labels, denoted as r, is calculated based on a 11-fold cross-validation classification procedure.The 11 fold cross-validation is based on the data matrix S, which is separated into the training and testing data sets for classification. Specifically, this cross-validation is only for calculating the prediction accuracy r. Can anyone give me some suggestion to do this? Thanks a lot!

is there any tutorial for me to read something about how to use libsvm? — user3761566, Aug 09 '14 at 13:46
yes it is, you need to install the library and you can use SVM with it.. — lakshmen, Aug 09 '14 at 14:31
I googled it for you, i'm kind like that. http://www.mathworks.co.uk/help/stats/classificationsvm.crossval.html — QED, Aug 09 '14 at 16:42
when I use this:svmModel = svmtrain(meas(trainIdx,:),group(trainIdx),'Autoscale',true,... the error shows: "undefined variable meas" How to fix this problem? — user3761566, Aug 09 '14 at 18:52

score 0 · Answer 1 · edited May 23 '17 at 11:49

0

To carry out SVM classification, use libsvm.

A good tutorial to start would be: https://sites.google.com/site/kittipat/libsvm_matlab

Matlab code for reference on cross validation:

Example of 10-fold SVM classification in MATLAB

Retraining after Cross Validation with libsvm

Installing LibSVM on MATLAB: http://www.youtube.com/watch?v=Wz_4h_bH7-c

edited May 23 '17 at 11:49

Community

1
1

answered Aug 09 '14 at 14:22

lakshmen

28,346
66
178
276

Building with 'Microsoft Visual C++ 2013 Professional'. Could Not Find C:\D disk\MATLAB\R2014a\libsvm-3.18\matlab\svmtrain.exp Could Not Find C:\D disk\MATLAB\R2014a\libsvm-3.18\matlab\svmtrain.exp MEX completed successfully. – user3761566 Aug 09 '14 at 15:44
it cannot find svmtrain.exp – user3761566 Aug 09 '14 at 15:45
When I try to load s_matrix groups = ismember(s_matrix,'setosa'); it shows error like this:Error using horzcat Dimensions of matrices being concatenated are not consistent. – user3761566 Aug 09 '14 at 16:46
How to fix this? I follow the steps on this website:http://stackoverflow.com/questions/3070789/example-of-10-fold-svm-classification-in-matlab/3071938#3071938 – user3761566 Aug 09 '14 at 16:49
when I use this:svmModel = svmtrain(meas(trainIdx,:),group(trainIdx),'Autoscale',true,... the error shows: "undefined variable meas" How to fix this problem? – user3761566 Aug 09 '14 at 18:51
@ Parag S. Chandakkar,I still do not know how to code this question by following an example in this link: http://stackoverflow.com/questions/25222661/how-to-fix-the-fisheriris-cross-classification?noredirect=1#comment39289610_25222661 – user3761566 Aug 10 '14 at 00:07

How to do cross validation SVM classifier

1 Answers1

Linked