load irisdataset.txt DataSet=irisdataset;
Dim=size(DataSet);
%----------------------- Selecting 3 random centres ----------------------% Selection=rand(1,3); Selection=Selection*Dim(1,1); Selection=ceil(Selection); %Selecting the row number.
%--------------------------- 3 random centres ----------------------------% Centre1=DataSet(Selection(1),:); Centre2=DataSet(Selection(2),:); Centre3=DataSet(Selection(3),:);
n=input('No Of Iterations : ') %----------------------- Partitional Algorithm ---------------------------% %-------------------------- K means Algorithm ----------------------------% for j=1:1:n count1=0; Mean1=zeros(1,4); count2=0; group1=[]; Mean2=zeros(1,4); group2=[]; count3=0; group3=[]; Mean3=zeros(1,4);
%Finding the minimum distance for i=1:1:Dim(1,1)
Pattern1(i)=sqrt((Centre1(1,1)-DataSet(i,1))^2+(Centre1(1,2)-DataSet(i,2))^2+(Centre1(1,3)-DataSet(i,3))^2+(Centre1(1,4)-DataSet(i,4))^2); Pattern2(i)=sqrt((Centre2(1,1)-DataSet(i,1))^2+(Centre2(1,2)-DataSet(i,2))^2+(Centre2(1,3)-DataSet(i,3))^2+(Centre1(1,4)-DataSet(i,4))^2); Pattern3(i)=sqrt((Centre3(1,1)-DataSet(i,1))^2+(Centre3(1,2)-DataSet(i,2))^2+(Centre3(1,3)-DataSet(i,3))^2+(Centre1(1,4)-DataSet(i,4))^2);
LessDist=[Pattern1(i) Pattern2(i) Pattern3(i)]; Minimum=min(LessDist);
%Finding the new centre if (Minimum==Pattern1(i)) count1=count1+1; Mean1=Mean1+DataSet(i,:); group1=[group1 i]; else if (Minimum==Pattern2(i)) count2=count2+1; Mean2=Mean2+DataSet(i,:); group2=[group2 i]; else count3=count3+1; Mean3=Mean3+DataSet(i,:); group3=[group3 i]; end end
end
%----------------------------- New Centres -------------------------------% Centre1=Mean1/count1; Centre2=Mean2/count2; Centre3=Mean3/count3; plot(j,count1,'r'); hold on plot(j,count2,'g'); plot(j,count3,'b');
end
hold off % specify the indexed color for each point icolor = ceil((DataSet(:,4)/max(DataSet(:,4)))*256);
figure, scatter3(DataSet(:,1),DataSet(:,2),DataSet(:,3),DataSet(:,4),icolor,'filled'); figure,
scatter3(DataSet(:,1),DataSet(:,2),DataSet(:,3),[],DataSet(:,4),'filled')
