Please use this identifier to cite or link to this item:
|Title:||A hybrid algorithm for k-medoid clustering of large data sets|
|Keywords:||Biology computing;Computational complexity;Data structures;Genetic algorithms;Pattern clustering;Search problems;Very large databases|
|Citation:||IEEE Evolutionary Computation, 2004. CEC Jun 2004|
|Abstract:||In this paper, we propose a novel local search heuristic and then hybridize it with a genetic algorithm for k-medoid clustering of large data sets, which is an NP-hard optimization problem. The local search heuristic selects k-medoids from the data set and tries to efficiently minimize the total dissimilarity within each cluster. In order to deal with the local optimality, the local search heuristic is hybridized with a genetic algorithm and then the Hybrid K-medoid Algorithm (HKA) is proposed. Our experiments show that, compared with previous genetic algorithm based k-medoid clustering approaches - GCA and RAR/sub w/GA, HKA can provide better clustering solutions and do so more efficiently. Experiments use two gene expression data sets, which may involve large noise components.|
|Appears in Collections:||Computer Science|
Dept of Computer Science Research Papers
Items in BURA are protected by copyright, with all rights reserved, unless otherwise indicated.