Clustering of the self-organizing map

The self-organizing map (SOM) is an excellent tool in exploratory phase of data mining. It projects input space on prototypes of a low-dimensional reg ular grid that can be effectively utilized to visualize and explore propert ies of the data. When the number of SOM units is large, to facilitate quant itative analysis of the map and the data, similar units need to be grouped, i.e., clustered. In this paper, different approaches to clustering of the SOM are considered, In particular, the use of hierarchical agglomerative cl ustering and partitive clustering using Ic-means are investigated. The two- stage procedure-first using SOM to produce the prototypes that are then clu stered in the second stage-is found to perform well when compared with dire ct clustering of the data and to reduce the computation time.