I was wondering if it is possible to cluster SOM results which is generally known as two-stage clustering? Some literature on the internet shows considerable promise such as the DS2L-SOM: http://cdn.intechopen.com/pdfs-wm/10455.pdf
. SOM is attractive because it offers reduction of the data dimensions, data analysis tools and is robust to null values and skewed data. Clustering the results of a large SOM is attractive for further identifying groups without prior knowledge of the number of clusters and those groups, which can then be investigated more qualitatively than standard SOM. Any advice on the recommended approach would be helpful. I was not sure that the SOM within RapidMiner can output the distance and density information required, to be input into a clustering method?