,,,-.YUNenghai,。andWANGLei,(-munication,UniversityofScienceandTechnologyofChina,Hefei23000%China)(,UniversityofScienceandTechnologyofChinaHe,e23002%China)(,ninUniversity,Tianjin300072,China)Abstract——Proposeanewensemblelearningalgo-rithmbasedonFisherlineardiscriminantanalysisfFLDorLDA、—classclassificationprob-,ordingtothesampleweights,andFLDisemployedineachsubsettoconstructanumberofbootstrapFLDsub-—nearestneighbor(KNN)pseudo-·—worlddatasets,wedemonstratethatouralgorithmissuperiortoothertechniques,——Fisherlineardiscriminantanalysis,En-semblelearning,,Bagging,Bootstrapsub-,suchasfacerecognitiontasks,featureextractiontechniqueisusedtoselectdiscriminantfea-turesoftheoriginaldataforseparationandclassification,whilereducingthedatadimensiontoeliminatedataredun—dancy.
Boosting Bootstrap FLD Subspaces in Classification.pdf 来自淘豆网m.daumloan.com转载请标明出处.
