（共56页PPT）UCAS模式识别SVM.pptxVIP

PatternRecognitionSupportVectorMachine(SVM)LiangWangCenterforResearchonIntelligentPerceptionandputingInstituteofAutomation,ChineseAcademyofSciences

Outline-IntroductiontoSVM:Historyandmotivation-Problemdefinition-SVM:TheLinearseparablecase-SVM:NonLinearseparablecase:-ThekernelTrick:discussiononKernelfunctions.-Softmargin:introducingtheslackvariablesanddiscussingthetrade-offparameter“C”.-ProcedureforchoosinganSVMmodelthatbestfitsourproblem(“K-fold”).-SomeApplicationsofSVM.-Conclusion:TheAdvantagesandDrawbacksofSVM.-