Mining a Growing Feature Map by Data Skeleton Modelling
Damminda Alahakoon, Saman Kumara Halgamuge, Balakrishnan Srinivasan · Studies in fuzziness and soft computing · 2001
The Growing Self Organising Map (GSOM) has been presented as an extended version of the Self Organising Map (SOM) which has significant advantages for knowledge discovery applications. In this article, we present a further extension to the GSOM in which the cluster identification process can be automated. The self-generating ability of the GSOM is used to identify the paths along which the GSOM grew, and these paths are used to develop a skeleton of the data set. Such a skeleton is then used as a base for separating the clusters in the data