lms_training.m

基于最小误差平方算法的线性分类程序

function W=Training(Sample1,Sample2);
[N1,l]=size(Sample1);
[N2,l]=size(Sample2);
N=N1+N2;
Y=zeros(N,1);
W=zeros(l+1,1);
S1=ones(l+1,N1);
S1(1:2,:)=Sample1';
S2=ones(l+1,N2);
S2(1:2,:)=Sample2';
X=[S1'; S2'];
Y(1:N1)=1;
Y(N1+1:N)=-1;
W=inv(X'*X)*X'*Y;