K-means++ Algorithm | MATLAB

%Code:

clc

clear all

close all

z=["Rob",27,70000;

"Michael",29,90000;

"Mohan",29,61000;

"Ismail",28,60000;

"Kory",42,150000;

"Gautam",39,155000;

"David",41,160000;

"Andrea",38,162000;

"Brad",36,156000;

"Angelina",35,130000;

"Donald",37,137000;

"Tom",26,45000;

"Arnold",27,48000;

"Jared",28,51000;

"Stark",29,49500;

"Ranbir",32,53000;

"Dipika",40,65000;

"Priyanka",41,63000;

"Nick",43,64000;

"Alia",39,80000;

"Sid",41,82000;

"Abdul",39,58000];

x=double(z(:,2));

y=double(z(:,3));

plot(x,y,'b.','linewidth',2);

xlabel('Age');

ylabel('Income');

title('Income Dataset');

hold on;

m=[];

k=input('Enter the k value of k means:');

pos = randi(length(x));

startx=x(pos);

starty=y(pos);

center=[startx starty];

d=[];

r=1;

while(r~=k)

for i=1:length(x)

    g=[x(i) y(i)];

    ka=dsearchn(center,g);

    nearestx=center(ka,1);

    nearesty=center(ka,2);

    distance=sqrt((nearestx-g(1))^2+(nearestx-g(1))^2);

    d=[d distance];

end

[e s]=max(d);

center=[center;[x(s) y(s)]];

x(s)=[];

y(s)=[];

[r c]=size(center);

d=[];

end

disp(center);

plot(center(:,1),center(:,2),'ko','linewidth',3);

grid on;

hold off;

x=double(z(:,2));

y=double(z(:,3));

cx=center(:,1);

cy=center(:,2);

mean_oldx=cx;

mean_newx=cx;

mean_oldy=cy;

mean_newy=cy;

outputx=cell(k,1);

outputy=cell(k,1);

temp=0;

while(temp==0)

    mean_oldx=mean_newx;

    mean_oldy=mean_newy;

    for ij=1:length(x)

        mina=[];

        mu=x(ij);

        nu=y(ij);

     for mk=1:length(cx)

         mina=[mina sqrt((mu-cx(mk))^2+(nu-cy(mk))^2)];

     end

     [gc index]=min(mina);

     outputx{index}=[outputx{index} mu];

     outputy{index}=[outputy{index} nu];

    end

    gmckx=[];

    gmcky=[];

    for i=1:k

        gmckx=[gmckx mean(outputx{i})];

        gmcky=[gmcky mean(outputy{i})];

    end

    cx=gmckx;

    cy=gmcky;

    mean_newx=cx;

        mean_newy=cy;

        gum=0;

        bum=0;

    if(mean_newx==mean_oldx)

        gum=1;

    end

    if(mean_newy==mean_oldy)

        bum=1;

    end

    if(gum==1 && bum==1)

        temp=1;

    else

            outputx=cell(k,1);

            outputy=cell(k,1);

    end

   

end

celldisp(outputx);

celldisp(outputy);

figure;

for i=1:k

    x=outputx{i};

    y=outputy{i};

    if(i==1)

        plot(x,y,'r*','linewidth',3);

    elseif(i==2)

        plot(x,y,'bo','linewidth',3);

    else

        plot(x,y,'g+','linewidth',3);

    end

    hold on;

    grid on;

end

xlabel('Age');

ylabel('Income');

title('kmean clustering on Income data');