Re: [Wekalist] Number of Support Vectors for LIBSVM