TITLE

Novel Hybrid Clustering Optimization Algorithms Based on Plant Growth Simulation Algorithm

AUTHOR(S)
Tavakolian, Rozita; Charkari, Nasroollah Moghaddam
PUB. DATE
December 2011
SOURCE
Journal of Advanced Computer Science & Technology Research;2011, Vol. 1 Issue 2, p84
SOURCE TYPE
Academic Journal
DOC. TYPE
Article
ABSTRACT
Data clustering as one of the important data mining techniques is a fundamental and widely used method to achieve useful information about data. In face of the clustering problem, clustering methods still suffer from trapping in a local optimum and cannot often find global clusters. In general, many evolutionary algorithms have been used to optimize the result of K-means clustering methods. However, getting away local optimum has been unavoidable. Plant Growth Simulation Algorithm (PGSA) is one of the most recently introduced evolutionary algorithms which simulates the plant phototropism process. In order to overcome the shortcoming of the available clustering methods, this paper presents two novel and efficient hybrid clustering algorithms based on PGSA, named PGKC stands for Plant Growth simulation algorithm with K-means for Categorizing data and PGKGA stands for Plant Growth simulation algorithm and K-means with Genetic Algorithm. The first applies PGSA on K-means algorithm to find near global optimal clusters. And the second hybridize PGKC and genetic algorithm with an objective of combining the speed and simplicity of PGKC and the searching capability of genetic algorithm to achieve a better performance. No previous method of combining plant growth simulation algorithm and genetic algorithm in literature has yet been found. Experimental results on different datasets from the UCI Machine Learning Repository demonstrate that the proposed algorithms can find better clusters in comparison with other methods in literature such as classic K-means, genetic algorithm with K-means and SOM neural network with K-means.
ACCESSION #
74120995

 

Share

Read the Article

Courtesy of THE LIBRARY OF VIRGINIA

Sorry, but this item is not currently available from your library.

Try another library?
Sign out of this library

Other Topics