SEQUENTIAL FUZZY CLUSTERING BASED ON NEURO-FUZZY APPROACH
Yevgeniy V. Bodyanskiy, A. O. Deineko, Ya. V. Kutsenko · Radio Electronics Computer Science Control · 2016
An on-line neuro-fuzzy system for solving data stream fuzzy clustering task and its self-learning procedures based on T. Kohonen’s rule are proposed in the paper. The architecture of proposed system consists of seven information processing layers and represents the hybrid of the Wang-Mendel system and clustering selforganizing network. During a learning procedure in on-line mode, the proposed system tunes both its parameters and its architecture. For tuning of membershipfunctions parameters of neuro-fuzzy system the method based on competitive learning is proposed. The hybrid neuro-fuzzy system tunes its synaptic weights, centers and width parameters of membership functions. Software that implements the proposed hybrid neuro-fuzzy system’s architecture has been developed. A number of experiments has been held in order to research the proposed system’s properties. Experimental results have proved the fact that the proposed system could be used to solve a sequential stream clustering task. The proposed system provides computational simplicity. A distinguishing feature of the proposed system is that this system combine supervised learning and self-learning procedures.