Thesis (BS Computer Science) -- University of the Philippines Mindanao, 2010

Data clustering is an act of partitioning an unlabeled data set into groups of similar objects. Each group called a cluster consists of objects that are similar between themselves and dissimilar to object of others clusters. This project aims to find an alternative method of clustering continuous data set using hybrid method using two predefined methods. The involved algorithms in the study are Multi-Elitist Particle Swarm Optimization and Tabu Search. There are many cases that classical PSO and Tabu Search are combined and achieved an outstanding outcome. With this, the study improvised the hybrid method by using modified approach for each algorithm: Multi-Elitist for the classical PSO and searching modifications on the existing Tabu Search wherein the study use the idea of swapping points. After implementing and repetitive testing with 30 runs for each combination of parameter settings, the graph shows the comparison between hybrid algorithms and its counterpart as well as the hybrid MEPSO-TS against other existing hybrid method like PSO-TS. The results dictated the domination of the MEPSO-TS against other algorithms in terms of the solution quality. But, have failed to achieve optimal solution time in some cases of the Comparison. Thus, further analysis on how to deal with time optimization maintaining solution quality would fill in this study