0

I am a beginner to SVM which i have successfully implemented one-class classification.Now i want to know about multi-class classification which am very much confused about.

I went through How to do multi class classification using Support Vector Machines (SVM) which i want the exact same output but the link does not have a specific example using windows.If anyone can help me out with an example in windows for both “ONE-AGAINST-ONE”,”ONE-AGAINST-ALL” methods of multi-class classification

Thanks

Community
  • 1
  • 1
user1051536
  • 185
  • 2
  • 3
  • 9

2 Answers2

1

Using libLinear you will not be able to have a similar output because it cannot predict probabilities. You should use libSVM for that.

LibLinear does not support multi-class classification by default, but you can download this tool from the official site and it can do the job.

If you want multi-class probability estimate, you can take a look at this tool

alexandrekow
  • 1,927
  • 2
  • 22
  • 40
  • thanks for the answer.I tried using train -s0 and then predict -b 1 and it gives me probabilty for all the classes.Does this mean multi-class classification? – user1051536 Jul 09 '14 at 15:27
  • Well if it produces the output you expected, I would say yes. – alexandrekow Jul 09 '14 at 15:48
  • thank you very much.It saved my day.I have same question but in other format.have a look at http://stackoverflow.com/questions/24685255/multi-label-prediction-using-liblinear.if possible please help me – user1051536 Jul 11 '14 at 13:37
0

I think the other answer is talking about multi-label classification or something because liblinear does support multi-class classification by default (Choose -s to be between 0 and 7, there are 8 different modes https://github.com/cjlin1/liblinear/blob/master/README)

For the input, you can use the same as binary classification, but just set the label to be the index of the class (between 0 and (# of classes - 1)) instead of ±1.

Alex Li
  • 246
  • 4
  • 11