Procedure: Multi Objective Subspace Clustering algorithm |
Begin |
Apply preprocessing phase on data set D |
= initial population of relevant subspaces /* Popsize = */ |
While (the termination criterion is not satisfied) |
for i = 1 to // for each individual in subspace |
Randomly select chromosomes from the population |
Call subspace_update ( ) |
Compute objective functions for current chromosomes |
Apply crossover operation with probability |
Apply mutation operator with mutation probability |
Compute objective functions for new offsprings |
end for |
end while |
Select the best solution from population |
End |
Procedure subspace_update ( ) |
Choose cluster centers {s = 1,…, from data set generated |
by preprocessing phase |
Repeat |
Compute the membership matrix |
for i =1 to do |
for j = 1 to do // —number of cluster centers |
if dist(, ) < dist(, ) then |
= 0; |
else |
= 1; |
end if |
end for |
end for |
Compute the cluster center: |
; |
Compute cluster weights encoded in each chromosome using |
membership matrix and cluster center |
until convergence |