Discrete particle swarm optimization and EM hybrid approach for naive bayes clustering

Authors:
Jing-Hua Guan;Da-You Liu;Si-Pei Liu
Affiliations:
Sch. of Computer S&T, Jilin Univ., Changchun, China;Sch. of Computer S&T, Jilin Univ., Changchun, China;Sch. of Computer S&T, Jilin Univ., Changchun, China
Venue:
ICONIP'06 Proceedings of the 13th international conference on Neural Information Processing - Volume Part II
Year:
2006

Citing 6
Cited 0

Bayesian classification (AutoClass): theory and results

Advances in knowledge discovery and data mining
Bayesian Network Classifiers

Machine Learning - Special issue on learning with probabilistic representations
Learning Bayesian networks for clustering by means of constructive induction

Pattern Recognition Letters - Special issue on pattern recognition in practice VI
An improved Bayesian structural EM algorithm for learning Bayesian networks for clustering

Pattern Recognition Letters
Learning Recursive Bayesian Multinets for Data Clustering by Means of Constructive Induction

Machine Learning - Special issue: Unsupervised learning
The Bayesian structural EM algorithm

UAI'98 Proceedings of the Fourteenth conference on Uncertainty in artificial intelligence

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper presents an improved Naive Bayes algorithm for clustering. Many researchers search for parameter values from incomplete data using EM (Expectation Maximization) algorithm. It is well-known that EM approach has a drawback – local optimal solution, so we propose a novel hybrid algorithm of the DPSO (Discrete Particle Swarm Optimization) and the EM approach to improve the global search performance. We then apply the approach to 4 real-world data sets from UCI repository and compare the performance of clustering by the new algorithm with by EM algorithm. In the comparison, the hybrid DPSO+EM algorithm exhibits more effectively and outperforms the EM approach.