Hierarchical Representatives Clustering with Hybrid Approach

Clustering is a discoveringprocess of meaningfulintbrmationby groupingsimilar data into compactclusters. Mostof traditional clustering methodsare in favor of small datasets andhavedifficulties handling very large datasets. Theyare not adequateclustering methodsfor partitioning hugedatasets in data mining perspective. Wepropose a newclustering technique, HRC(hierarchicalrepresentatives clustering), that can be applied to large datasets andfind clusters withgood quality. HRCis a twophase algorithmthat take advantage of a hybrid approachthat combineSOMandhierarchical clustering. Experimentalresults showthat HRC candiscoverbetter clusters efficiently in comparisonto traditional clustering methods.

• Type: Conference
• Year: 2001
• Where: FLAIRS
• Authors: Byung-Joo An, Eunju Kim, Yillbyung Lee
• Keywords: Artificial Intelligence, FLAIRS 2001, Large Datasets, Mostof Traditional Clustering, Traditional Clustering