Support Vector Machine (SVM) is a popular machine learning technique proposed by Vapnik and co-workers in 1992. I know that several existing tools are available for SVM such as:

- MySVM
- SVMlight
- Torch
- LibSVM

I’m curious to know which of these toolboxes you have used and liked (for those of you who have done SVM). If you think at another toolbox, feel free to comment as well.

P.S. Since I have to go to the army, Data Mining Research will not be updated the following three weeks. I’m looking forward to see you in July.

Hello,

actually, for regression purposes (function approximation), I like to use the LS-SVMLab (http://www.esat.kuleuven.ac.be/sista/lssvmlab/) It implements least squares support vector machines, which I prefer over epsilon-svm or nu-svm, (but this is a personnal choice).

damien

or shogun, which comes with many SVM implementations …

I have used LibSVM for regression and one-class classification. Sometimes I have use the e1071 package for R, but it is basically the LibSVM toolbox.

Thanks for your comments. On my side, I’m using the Statistical Pattern Recognition Toolbox for Matlab and it is quite easy to use. However, I don’t know about its computational efficiency.