Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2006

Data clustering is a problem that deals with classification of objects within the data set into clusters such that items in the same cluster have a high degree of similarity. Known heuristic algorithms are applied to solve the problem. In this study, Particle Swarm Optimization (PSO) hybrid with Simulated Annealing (SA) was used to cluster data on Iris data set. PSO is relatively new family of algorithm, which is a population ?based stochastic optimization technique while SA is an algorithm, which is a population-based stochastic optimization technique while SA is an algorithm that concerns with finding global extremum of the function and works on a single solution. Different sets of parameter values were tested on the algorithm to determine which setting best suits the data. Results showed that smaller parameter values for SA and PSO parameters except of inertia weight performed significantly faster while larger parameter values of all parameter except inertia gave better solution quality. The result also showed that number of hits or assignment of data to a cluster is somewhat bad. However, PSO-SA algorithm is still a promising alternative to cluster data on Iris data set if further improvements can be done.