[Home]


TWSVC

A Matlab code for twin support vector clustering. (You could Right-Click [Some Results] , and Save, then you can download the results.)


Reference

Zhen Wang, Yuan-Hai Shao*, Nai-Yang Deng. TWSVC: twin support vector machine for clustering[J]. IEEE TNNLS. 2015


Main Function
function pY= IteOne(X,Y,c) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % TWSVC: one iteration in twin support vector clustering % % pY= IteOne(X,Y,c,mu) % % Input: % X: Data matrix. (nolinear version is obtained by K(X,X)) % Y: First initial labels of X (can be given randomly or by NN-graph). % then, it would be updated in iteration. % Parameters - c,mu. The fields in options that can be set: % c: (0,inf) Paramter to tune the weight. % % Output: % pY: The prediction of X in this iteration. % Examples: % X=rand(50,10); % Y = randint(50,1,[1,5]); % c=1;mu=0.1; % pY= IteOne(X,Y,c) % Reference: % Zhen Wang, Yuan-Hai Shao, Nai-Yang Deng, "Twin support vector machine for clustering" IEEE TNNLS 2015 % % Version 2.0 --Nov/2016 % % Written by Zhen Wang (wangz11@mails.jlu.edu.cn) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Main %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% %tic; tol=0.001; eps=0.0000001; num=max(Y); totalu=zeros(1+size(X,2),num); for i=1:num inputA=X(Y==i,:); inputB=X(Y~=i,:); u0=FirstStep(inputA); [m1,n]=size(inputA); m2=size(inputB,1); if mu<=0 u=zeros(n+1,1); end ite=0; som=1; con=0; while som>tol && ite<30 ite=ite+1; u=u0; e1=ones(m1,1); e2=ones(m2,1); G=[inputB,e2]; G=diag(sign(inputB*u(1:n,:)+u(n+1,1)))*G; H=[inputA,e1]; kerH=G*((H'*H+eps*eye(n+1))\G'); kerH=(kerH+kerH')/2; gamma=quadprog(kerH,-e2,[],[],[],[],0*e2,c*e2,[],optimset('display','off')); % gamma=qpSOR(kerH,0.7,c,0.05); %SOR u0=(H'*H+eps*eye(n+1))\G'*gamma; som=norm(u-u0); end totalu(:,i)=u0; end [tmp,pY]=min(abs([X,ones(size(X,1),1)]*totalu),[],2); end %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% % Additional functions %%%%%%%%%%%%%%%%%%%%%%%%%%%%%% function u=FirstStep(A) % compute: min ||Aw+be||, s.t. ||w||=1. % u=[w;b] m=size(A,1); H=A'*(1/m*ones(m,m)-eye(m))*A; [V,D]=eig(H); [tmp,n]=min(abs(diag(D))); w=V(:,n); b=-1/m*sum(A,1)*w; u=[w;b]; end
Contacts

Any question or advice please email to wangz11@mails.jlu.edu.cn.