Get the support vectors generated by SVM


I need help in the following issue: I want to perform kmeans cluster with k clusters. Then for each two clusters i would like to run SVM and get the support vectors to be stored in one table that similar to the original input file just keeping those supports vectors generated by SVM.

How i can perform that using k-means and svm node (SVM Learner or other similar nodes).




Hi Malik,

Could you please give some more details on what you are trying to achieve? 



Hi Roland

Let us make it simple. SVM solutions are support vectors of that data. Those vectors are actually part of the input vectors to SVM. How can I get those vectors?  If the data consists from vi , where i=1,n then the SVM find the support vectors that could be a number of vi. I need those vi.