A Novel Membrane Clustering Algorithm Based on Tissue-like P System

Abstract

Clustering is a process of partitioning data points into different clusters due to their similarity, as a powerful technique of data mining, clustering is widely used in many fields. Membrane computing is a computing model abstracting from the biological area, these computing systems are proved to be so powerful that they are equivalent with Turing machines. In this paper, a modified inversion particle swarm optimization was proposed, this method and the mutational mechanism of genetics algorithm were used to combine with the tissue-like P system, through these evolutionary algorithms and the P system, the idea of a novel membrane clustering algorithm could come true. Experiments were tested on six data sets, by comparing the clustering quality with the GA-K-means, PSO-K-means and K-means proved the superiority of our method.