Hierarchiczne aglomeracyjne sieci SOM w analizie skupień
The hierarchical agglomerative SOM in the cluster analysis

Summary/Abstract: Self-learning artificial neural networks type of SOM are one of the most effective data mining tools which are used in grouping and classification of multidimensionaldata. The decrease in network efficiency SOM clustering and classification of data oftenresults from the assumed redundant network structure and a significant increase of deadneurons in the network. The process of self-learning of the network becomes unnecessarilylong. One possibility of solving this problem is to build a hierarchical agglomerative SOM network. In these networks, there are two approaches: thematic and based on clusters. The aim of this paper is to analyze the properties of agglomerative HaSOM network in the cluster analysis.

• Details
• Contents