Clustering, Noise Reduction and Visualization Using Features Extracted from the Self-Organizing Map

摘要: This paper presents an analysis of a feature space generated by extracting properties related to pattern density and Euclidean distances between neurons from the self-organizing map network. Hence, along with weight vector, each neuron has 2-D vector associated it, whose components are extracted U-matrix hit matrix, where latter is based on hyperspheres centered neuron. collection vectors, that represents network, partitioned into different groups, their labels carried back data as well grid, in order perform tasks clustering, noise reduction visualization. Experiments were out using synthetic real world sets.