Improving the Accuracy and Efficiency of the k-means clusteringusing Optimization Algorithms

Clustering is a popular data analysis and data mining technique. The k-means clustering algorithm is one of themost commonly used methods. However, it highly depends on the initial solution and is easy to trap into the localoptimal. For overcoming these disadvantages of the k-means method in this paper a Swarm Clustering Algorithm isproposed based on the standard K-Means clustering algorithms, which are used as fitness functions for a SwarmIntelligence algorithm. The motivation is to exploit the search capability of Swarm Intelligence algorithms and to avoidthe major limitation of falling into locally optimal values of the K-Means algorithm. Because of the inherent parallelnature of the Swarm Intelligence algorithms, since the fitness function can be evaluated for each individual in anisolated manner, we have developed the parallel implementation, comparing the performances with their algorithm.Experiments with 2 bench-mark datasets have shown similar or slightly better quality of the results compared tostandard K-Means algorithm and other algorithm. the experiment results show that proposed algorithm clustering hasnot only higher accuracy but also higher level of stability. And the faster convergence speed can also be validated bystatistical results.

کلیدواژه ها:

clustering ، K-Means algorithm ، Optimization