Search

This paper focuses on automatic fuzzy clustering problem and proposes a novel automatic fuzzy clustering method that employs an extended membrane system with active membranes that has been designed as its computing framework. The extended membrane system has a dynamic membrane structure; since membranes can evolve, it is particularly suitable for processing the automatic fuzzy clustering problem. A modification of a differential evolution (DE) mechanism was developed as evolution rules for objects according to membrane structure and object communication mechanisms. Under the control of both the object's evolution-communication mechanism and the membrane evolution mechanism, the extended membrane system can effectively determine the most appropriate number of clusters as well as the corresponding optimal cluster centers. The proposed method was evaluated over 13 benchmark problems and was compared with four state-of-the-art automatic clustering methods, two recently developed clustering methods and six classification techniques. The comparison results demonstrate the superiority of the proposed method in terms of effectiveness and robustness. [ABSTRACT FROM AUTHOR]

Membrane computing is a class of distributed parallel computing models. Inspired from the structure and inherent mechanism of membrane computing, a membrane clustering algorithm is proposed to deal with automatic clustering problem, in which a tissue-like membrane system with fully connected structure is designed as its computing framework. Moreover, based on its special structure and inherent mechanism, an improved velocity-position model is developed as evolution rules. Under the control of evolution-communication mechanism, the tissue-like membrane system cannot only find the most appropriate number of clusters but else determine a good clustering partitioning for a data set. Six benchmark data sets are used to evaluate the proposed membrane clustering algorithm. Experiment results show that the proposed algorithm is superior or competitive to three state-and-the-art automatic clustering algorithms recently reported in the literature. [ABSTRACT FROM AUTHOR]