Material covered

  • Primal SVM demo in scilab, separable and inseparable data
  • Turning the SVM QP into a dual QP
  • Dual SVM demo in scilab, demonstration of support vectors
  • Sequential minimal optimization (SMO) algorithm
  • The kernel trick, common kernels used
  • Kernel properties and embedding
  • Dual SVM with nonlinear kernels -- scilab demo

Last modified: Friday, 31 August 2007, 10:05 AM