Muhammad Nadzeri Munawar
31 linhas
772 B
Matlab
31 linhas
772 B
Matlab
|
|
function [results] = kNN(TrainingSet,GroupTrain,TestSet,k)
|
|
[L,L1] = size(TestSet);
|
|
results = zeros(L);
|
|
for i = 1:L
|
|
[L2,L3] = size(TrainingSet);
|
|
data_distance = zeros(L2);
|
|
for j = 1:L2
|
|
p = TrainingSet(j,:);
|
|
q = TestSet(i,:);
|
|
|
|
data_distance (j) = sum(abs(p-q)./(abs(p) + abs(q)));
|
|
end
|
|
|
|
[B,I] = sort(data_distance);
|
|
|
|
nearest = GroupTrain(I(1:k));
|
|
nearest_distance = B(1:k);
|
|
sse = nearest_distance(1)/nearest_distance(2);
|
|
% disp(sse);
|
|
% disp(nearest);
|
|
if sse<=0.97
|
|
results(i)=nearest(1);
|
|
else
|
|
results(i)=mode(nearest);
|
|
end
|
|
%results(i)=mode(nearest);
|
|
end
|
|
end
|
|
|