A hybrid approach to learn Bayesian networks using evolutionary programming
Document Type
Conference paper
Source Publication
Proceedings of the 2002 Congress on Evolutionary Computation, CEC 2002
Publication Date
1-1-2002
Volume
2
First Page
1314
Last Page
1319
Publisher
IEEE Computer Society
Abstract
A novel hybrid framework is reported that improves upon our previous work, MDLEP, which uses evolutionary programming to solve the difficult Bayesian network learning problem. A new merge operator is also introduced that further enhances the efficiency. As experimental results suggest, our hybrid approach performs significantly better than MDLEP.
DOI
10.1109/CEC.2002.1004433
Publisher Statement
Copyright © 2002 IEEE. Access to external full text or publisher's version may require subscription.
Additional Information
Paper presented at the IEEE World Congress on Computational Intelligence (WCCI2002), May 12-17, 2002, Honolulu, Hawaii.
ISBN of the source publication: 9780780372825
Full-text Version
Publisher’s Version
Language
English