Abstract. High quality clustering is impossible without a priori information about clustering criteria. The paper describes the development of new clustering technique based on chaotic neural networks that overcomes the indeterminacy about number and topology of clusters. Proposed method of weights computation via Delaunay triangulation allows to cut down computing complexity of chaotic neural network clustering.

Abstract. This paper proposes an improved model of chaotic neural network used to cluster high-dimensional datasets with cross sections in the feature space. A thorough study was designed to elucidate the possible behavior of hundreds interacting chaotic oscillators. New synchronization type -fragmentary synchronization within cluster elements dynamics was found. The paper describes a method for detecting fragmentary synchronization and it's advantages when applied to data mining problem.

Abstract. The paper describes a unified approach to solve clustering and classification problems by means of oscillatory neural networks with chaotic dynamics. It is discovered that self-synchronized clusters once formed can be applied to classify objects. The advantages of distributed clusters formation in comparison to centers of clusters estimation are demonstrated. New approach to clustering on-the-fly is proposed.

No abstract

No abstract