A Novel Framework for Discovering Robust Cluster Results

We propose a novel method, called heterogeneous clustering ensemble (HCE), to generate robust clustering results that combine multiple partitions (clusters) derived from various clustering algorithms. The proposed method combines partitions of various clustering algorithms by means of newly-proposed the selection and the crossover operation of the genetic algorithm (GA) during the evolutionary process.

• Type: Conference
• Year: 2006
• Where: DIS
• Authors: Hye-Sung Yoon, Sang-Ho Lee, Sung-Bum Cho, Ju Han Kim
• Keywords: DIS 2006, Heterogeneous Clustering, Method Combines Partitions, Theoretical Computer Science, Various Clustering Algorithms