KNN的matlab简单实现代码

function target=KNN(in,out,test,k)
% in: training samples data,n*d matrix
% out: training samples' class label,n*1
% test: testing data
% target: class label given by knn
% k: the number of neighbors
ClassLabel=unique(out);
c=length(ClassLabel);
n=size(in,1);
% target=zeros(size(test,1),1);
dist=zeros(size(in,1),1);
for j=1:size(test,1)
cnt=zeros(c,1);
for i=1:n
dist(i)=norm(in(i,:)-test(j,:));
end
[d,index]=sort(dist);
for i=1:k
ind=find(ClassLabel==out(index(i)));
cnt(ind)=cnt(ind)+1;
end
[m,ind]=max(cnt);
target(j)=ClassLabel(ind);
end