Abstract
Cluster analysis is a mult ivariate analysis technique used to classify objects such that the objects in a cluster are very similar and the objects in different clusters are quite different. This study will discuss the non-hierarchical clustering methods. The methods are C-Means Cluster and Fuzzy C-Means Cluster. These methods are suitable for large data and continuous variables. This study would also present the application of the methods on the case of village grouping according to the underdevelopment status in two regions of level II (Kota Metro and Kabupaten Lampung Timur) in Lampung Province. The unit of observations in this study are 257 villages in Kota Metro (22 villages) and Kabupaten Lampung Timur in Lampung Province obtained from the Village Potential Statistics (Podes - Potensi Desa) 2008.
The results show that the optimal cluster in Kota Lampung data is 4, with a minimum value of the Fukuyama-Sugeno validity index is at -45.4649. As for the data of Kabupaten Lampung Timur, theoptimumnumber ofclustersis13,with aminimum value of the Fukuyama-Sugeno validity index is at 196.9629.
