Least Squares SVM Ensemble viaDiversity(Relationship) LearningHua Qiang
Least Squares SVM Ensemble via Diversity(Relationship) Learning Hua Qiang
ContentsIntroduction中Notation-Ensemble via Diversity LearningDLeast Squares SvM+Model & Optimization中Experiment
Contents Introduction Least Squares SVM Notation Ensemble via Diversity Learning Experiment Model & Optimization
LS-SVM Ensemble via Diversity Leaning-Introduction Ensemble methods, whichtrainmultiplelearners for a set of data. The diversity of the component learners hasbeen recognized as a key to a good ensemble
• The diversity of the component learners has been recognized as a key to a good ensemble. LS-SVM Ensemble via Diversity Leaning - Introduction • Ensemble methods, which train multiple learners for a set of data
LS-SVM Ensemble via Diversity Leaning-Notation(L,) the set of K learnersX=(x,X2.., X~)- the training setX, ERdy = (yi,..., yn) - the corresponding labelsy, E(-1,+1)the linear function for L,:f,(x) = w, x +b
— the set of K learners 1 2 x x x x { , ,., } = N 1 y { , , }n = y y — the training set — the corresponding labels 1 { }K Li i= the linear function for Li : LS-SVM Ensemble via Diversity Leaning - Notation ( ) x w xT i i i f b = + x d j R y { 1, 1} j − +
LS-SVM Ensemble via Diversity Leaning- Ensemble via Diversity Learningdiv(wi, W,KNZZL(y,w,x,)minWi...Wki=l j=l[w, l, ≤0 (Vie[K])s.t.div,((w,...Wk)) ≥qRef. :Yang Yu, Yu-Feng Li and Zhi-Hua Zhou, Diversity Regularized Machine,ijcai2011
LS-SVM Ensemble via Diversity Leaning - Ensemble via Diversity Learning 1 2 1 2 1 2 div( , ) 1 w w w w w w T = − 1 ,., 1 1 2 1 min ( , ) . . ( [ ]) ({ ,., }) w w w x w w w K K N T j i j i j i p K L y s t i K div q = = Ref. :Yang Yu, Yu-Feng Li and Zhi-Hua Zhou, Diversity Regularized Machine, ijcai2011