Abstract
Describes an original method: the Bayesian iterative method of clustering on the basis of feature selection. The comparison of clustering results obtained using kazimierzowska method with the results of applying the classical k-means. As clusterization used neural network self-organizing maps (SOM), i.e. of the neural network Kohonen. Based on real data applied neural network model for clustering with individual in-depth assessment of the solvency of enterprises – taxpayers
Keywords
neural networks, Kohonen self-organizing maps, kazimierzowski method, k - means, clustering, synaptic weight of the topological neighborhood
Comments
No posts found