Data clustering using virtual population based incremental learning algorithm with similarity matrix encoding strategy

ABSTRACT

Data clustering is a good benchmark problem for testing the performance of many combinatory optimization methods. However, very few works have been done on using the estimation of distribution algorithms for solving the problem of data clustering. The purpose of this paper is to demonstrate the effectiveness of the estimation of distribution algorithms for solving the problem of data clustering. In particular, a novel encoding strategy termed as the Similarity Matrix Encoding strategy (SME) and a Virtual Population Based Incremental Learning algorithm using SME encoding strategy (VPBIL-SME) are proposed for clustering a set of unlabeled instances into groups. Effectiveness of VPBIL-SME is confirmed by experimental results on several real data sets.

REFERENCES

• 1: Y. Hong, S. Kwong, Q. Ren, and X. Wang. A comprehensive comparison between real population based tournament selection and virtual population based tournament selection. In IEEE Congress on Evolutionary Computation (CEC2007), pages 445--452, 2007.

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.5 PATTERN RECOGNITION
      I.5.3 Clustering
          Subjects: Algorithms

General Terms:
Algorithms

Keywords:
data clustering, similarity matrix encoding strategy, virtual population based incremental learning algorithm