lec3

  The Support Ve ct or Machi ne So fa r we have used a reference as sumpt ion tha t there exists a linear classifier that has a larg e ge ometric margin, i.e., whose decision b oundar y is we ll separa ted from all the training images (e xamples). Such a larg e margin classifier seems like one we would like to use. C an’t we find it more directly? Yes, we can. The clas sifier is known as the Support Ve ctor M achine or SVM for short.