-
Notifications
You must be signed in to change notification settings - Fork 0
/
Assignment.m
67 lines (51 loc) · 2.33 KB
/
Assignment.m
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
function [ dataset, distortion ] = Assignment( dataset,centroids, K ,colors,attempt)
%UNTITLED Summary of this function goes here
% Detailed explanation goes here
% size/rows of centroid matrix == K
% attempt is how many times we updated our mean.
for i=1:size(dataset(:,1),1)
for j=1:size(centroids(:,1),1)
% add them in a matrix to calculate euc distance
X = [dataset(i,1) dataset(i,2);centroids(j,1),centroids(j,2)];
% calculating the euc dist b/w centroid and the point x_n and
% adding it to the fourth column of centroids matrix.
centroids(j,4) = pdist(X);
% here j is representing the ID/class of K.Adding to the 5th
% column of the centroids matrix.
centroids(j,5) = j;
end
% get the row with min euc distance of centroid to that point.4th
% column represents the euc distances calculated for that
% particular point from all the centroids.
minEuc = centroids(:,4);
minClass = centroids(:,5);
%assign the class to the datapoint for being closest to that class.
%By class I mean K or centroid.5th column represents the class/id
%of the centroid.
%row and column containing the min euclidean distance
[r,c] = find(minEuc==min(min(minEuc)));
dataset(i,4) = min(minEuc); % euc dist
dataset(i,5) = minClass(r,1); % class
end
clf('reset');
figure(1);
hold on;
distortion = 0;
for i=1:K
classpoints = dataset(dataset(:,5)==i,:);
scatter(classpoints(:,1),classpoints(:,2),20,colors(i,:));
scatter(centroids(i,1),centroids(i,2),50,'b+');
% here 'i' is the loop iteration or class.
% classpoints contains all the points with class value 'i'.
x = [classpoints(:,1),classpoints(:,2)];
% centroid number 'i'.
y = [centroids(i,1),centroids(i,2)];
% remove the row with value 0 since that is the euc distance of centroid to itself.
% after removing that you can select the value with minimum distortion.
JFinder = pdist2(x,y);
[r,c] = find(JFinder==min(min(JFinder)));
JFinder(r,:) = [];
distortion = distortion + min(JFinder);
drawnow;
end
end