Page 91 - Algorithm Collections for Digital Signal Processing Applications using MATLAB

2. Probability and Random Process 79

for i=1:1:6
[x,y]=find(CLUSTERNO==i);
col=[];
for k=1:1:length(x)
col= [col a(x(k),y(k))];
end
plot(col,zeros(1,length(col)),strcat(u(i),'*'))
hold on
b(i)=mean(col);
end
pause(0.01)
end

________________________________________________________________________

4. FUZZY K-MEANS ALGORITHM FOR PATTERN
RECOGNITION

Consider the problem described in the section 3 for classifying the
normalized marks into 6 clusters for the assignment of grades.
In fuzzy k-means technique, fuzzy set theory is used to obtain the optimal
values of the centroid.
In this technique the particular vector (In this problem it is the
normalized mark scored by the student) belongs to all the 6 clusters with
different membership values. For instant the vector 0.3 belongs to the
different clusters with different membership values as given below

Cluster 1 = {0.3 (0.0001)}
Cluster 2 = {0.3 (0.0448)}
Cluster 3 = {0.3 (0.0022)}
Cluster 4 = {0.3 (0.0031)}
Cluster 5 = {0.3 (0.9492)}
Cluster 6 = {0.3 (0.0007)}

The numbers in bold letters are the corresponding membership values.
(i.e.) 0.3 belongs to the cluster 1 with membership value 0.0001 and belongs
to the cluster 2 with membership value 0.0448 and so on. Note that sum of
the membership values is 1.