close all; clear; clc;
% dimensions
k = 3;
% number of clusters
dim = 4;
% dimension of data
% read dataset from file
file = fopen('data.txt');
data = fscanf(file,'%f');
data = reshape(data,dim,[]);
data = data';
n = size(data,1);
fclose(file);

% initialize centroids
rp = randperm(n);
ct = data(rp(1:k),:);
inict = ct;

g = zeros(k,n); % cluster data
iter = 0;       % clustering
while true
% to calculate new clusters
    ctn = zeros(k,dim);
    gn = zeros(k,n);

       for i = 1:n
           % for each data
           % selecting cluster
           d = zeros(k,2);

            for j = 1:k
                % for each centroid
                % cluster index
                d(j,1) = j;
                % distance to centroid
                d(j,2) = sqrt( sum( ( ct(j,:)-data(i,:) ).^2 ) );
            end
             d = sortrows(d,2);
             % sort by distance
             c = d(1,1);
             % selected cluster
             gn(c,i) = 1;
             % classify
             ctn(c,:) = ctn(c,:) + data(i,:);
         end