Novel K-means Clustering Algorithm Combining Particle Swarm Optimization and Bacterial Foraging

This paper, we propose a novel clustering algorithm combining bacterial foraging algorithm (BFO) which has high global search ability with particle swarm algorithm (PSO)which has high local search ability based on the initial values of traditional k-means clustering algorithm is sensitive and easy to fall into local optimum solution, in order to optimize the initial cluster centers of K-means clustering algorithm. The chemotaxis of bacteria is simplified as the process of searching for the optimal solution of the particles in the particle swarm, and then the bacterias complete the further operation of reproduction and elimination-dispersal. The optimal solution of the hybrid algorithm is determined as the initial clustering center, and the disadvantages of the k-means algorithm are solved. The results on standard date set Iris, Wine, Glass of UCI display that the accuracy and stability of this proposed algorithm is higher than that of the popular clustering algorithms, at the same time, it can solve complex optimization problems most effectively.