ı want use libsvm for classifying my test dataset but ı can't. I went to "File > Set Path > Add folder > [Select the LBSVM folder] > Save . And ı run this code: SVMStruct = svmtrain(histtrain,grupp); snc = svmclassify(SVMStruct,histogramtst) But matlab gave this error: SVMTRAIN only supports classification into two groups. GROUP contains 10 groups. Can you help me ı have 405 images for train dataset and ı have 192 images for 192 images what can ı do
Hi,
in principle the SVM only supports a two class problem (binary classification). If you have more classes to classify you need different strategies, namely, e.g., either you have to build (i) a binary classifier with one group to the rest (one-to-all) or (ii) you have to distinguish between each pair of groups (one-versus-one). For (i) you can use a winner-takes-all strategy and for (ii) you can use a max-win strategy.
At http://www.mathworks.com/matlabcentral/fileexchange/39352-multi-class-svm I've found an example for a multiclass SVM, but I have not tested it, therefore I'm not shure if it is bug free.
good luck!
thanks for your reply. ı have tested this link http://www.mathworks.com/matlabcentral/fileexchange/39352-multi-class-svm
But result is always same for example ı have 10 class and all my test results is first class. This code gives always same result:(
ok, have you tried this out: http://www.csie.ntu.edu.tw/~cjlin/libsvm/faq.html#f808
LibSVM itself supports multiclass classification, and for this there exists a Matlab interface http://www.csie.ntu.edu.tw/~cjlin/libsvmtools/ovr_multiclass/
thanks very much for your reply but ı dont understand this code for example which is testfeature,trainfeature and class labels for y,x,cmd
- Badges
- Science topic