Improvement of CB & BC Algorithms (CB* Algorithm) for Learning Structure of Bayesian Networks as Classifier in Data Mining

Benhard Sitohang, G. A. Putri Saptawati

Abstract


There are two categories of well-known approach (as basic principle of classification process) for learning structure of Bayesian Network (BN) in data mining (DM): scoring-based and constraint-based algorithms. Inspired by those approaches, we present a new CB* algorithm that is developed by considering four related algorithms: K2, PC, CB, and BC. The improvement obtained by our algorithm is derived from the strength of its primitives in the process of learning structure of BN. Specifically, CB* algorithm is appropriate for incomplete databases (having missing value), and without any prior information about node ordering.

Full Text:

PDF

References


Cheng, J. & Geiner, R., Comparing Bayesian Network Classifiers, Proceedings of 15th International Conference on Uncertainty in AI, 1999.

Friedman, N. & Koller, D., Learning Bayesian Networks from Data, 1998.

Steck, H. & Tresp, V., Bayesian Belief Networks for Data Mining, Siemens AG, Corporate Technology Information and Telecommunications, Munich, Germany.

Madden, M.G., A New Bayesian Network Structure for Classification Tasks.

Heckerman, D., A Tutorial on Learning Bayesian Network, Technical Reports, 1995.

Jensen, F.V., Bayesian Networks and Decision Graphs, Springer, 2001.

Cheng, Jie & Greiner R., Learning Bayesian Network Classifier: Algorithms dan System, Proceeding of 14th Biennial conference of the Canadian Society for Computational Studies of Intelligence, 2001.

Neapolitan, R.E., Learning Bayesian Networks, Pearson Prentice Hall, 2004.

Han, J. & Kamber, M., Data Mining: Concepts and Techniques, Morgan Kaufman, 2001.

Kruse, Rudolf & Borgelt, C., Graphical Models: Analysis Tool for Data Mining, Dept. of Knowledge Processing and Language Engineering, Otto-von-Guericke-University of Magdeburg, Germany, 2000.

Cheng, J., Bell, D. & Liu, W., Learning Bayesian Networks from Data: A Efficient Approach Based on Information Theory, 1997.

Heckerman, Chickering, A Comparison of Scientific and Engineering Criteria for Bayesian Model Selection, Microsoft Research, 1997.

Hirsalmi, M., Method Feasibility Study: Bayesian Networks, Research Report, 2000.

Murphy, Kevin P., A Brief Introduction to Graphical Model and Bayesian Networks, UC Berkeley, 2003.

Pearl, J. Graphical Models for Probabilistic and Causal Reasoning, Computer Science Department, University of California, 1997.

Sophia, A., Algoritma PC sebagai Alternatif Pendekatan Analisis Dependensi untuk Konstruksi Struktur Bayesian Network dalam Data Mining, Final Project Dept of Informatics, Institute of Technology Bandung, 2005.

Sandhyaduhita P.I., Algoritma CB: Algoritma yang Dibangun dengan Dua Pendekataan untuk Konstruksi Bayesian Network dalam Data Mining, Final Project Dept of Informatics, Institute of Technology Bandung, 2005.

Singh, M. & Valtorta, M., Construction of Bayesian Network Structures from Data: a Brief Survey and Efficient Algorithm, Department of Computer Science, Univ. of South Carolina, Columbia, USA, 2005.

Ramoni, M. & Sebastiani P., Learning Bayesian Networks from Incomplete Databases, KMI-TR-43, 1997.

Maharani, H., Konstruksi Struktur Bayesian Network dalam Data Mining untuk Basis Data Incomplete dengan Metode Bound and Collapse, Final Project Dept of Informatics, Institute of Technology Bandung, February 2005.




DOI: http://dx.doi.org/10.5614%2Fitbj.ict.2007.1.1.3

Refbacks

  • There are currently no refbacks.


Contact Information:

ITB Journal Publisher, LPPM – ITB, 

Center for Research and Community Services (CRCS) Building Floor 7th, 
Jl. Ganesha No. 10 Bandung 40132, Indonesia,

Tel. +62-22-86010080,

Fax.: +62-22-86010051;

e-mail: jictra@lppm.itb.ac.id.