Digital Library

cab1

 
Title:      SELF-TUNING CLUSTERING USING THE INFORMATION GEOMETRY TECHNIQUE
Author(s):      Renata Avros, Zeev Barzily, Zeev Volkovich
ISBN:      978-972-8939-93-9
Editors:      António Palma dos Reis and Ajith P. Abraham
Year:      2013
Edition:      Single
Keywords:      Clustering, Cluster validation, Distance learning.
Type:      Full Paper
First Page:      13
Last Page:      20
Language:      English
Cover:      cover          
Full Contents:      click to dowload Download
Paper Abstract:      In this paper a new adaptive clustering algorithm is presented. This method incorporates metric learning attitude based on the information geometry methodology and uses the pair-wise constraints created at each step in order to refine the clustering solution. Individual distance weight matrices of clusters are calculated at relying on the previously obtained information with reference to the desired partition. The weighed distance matrices obtained as a collateral outcome of the partitioning process are considered as a new cluster quality characteristic. Numerical experiments demonstrate the potential of such an approach in clustering and in predicting the “true” number of clusters.
   

Social Media Links

Search

Login