Dear All, I m trying to build SVM model with large dataset. However, svm from package e1071 works slowly (takes hours) to build model for like 1 million observations. I was thinking about CVM (Core Vector Machines: http://www.cse.ust.hk/~jamesk/papers/jmlr05.pdf<http://www.cse.ust.hk/%7Ejamesk/papers/jmlr05.pdf>), however couldn't find any R interface for this model. Does anybody know how to deal with SVM for large datasets. Thank you for any advices. -Alex [[alternative HTML version deleted]]