Abstract
This paper shows a new method of explaining Bayesian networks by creating descriptions of their properties in a manner closer to the human perceptual abilities, i.e., decision rules in the IF...THEN form (called by us belief rules). The conversion method is based on the cause and effect analysis of the Bayesian network quantitative component (the probability distribution). Proposed analysis of the quantitative component leads to a deeper insight into the structure of knowledge hidden in the analyzed data set.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
Notes
- 1.
The Bayesian network has been generated using BeliefSEEKER system [11]. The system generates learning models in a form of belief networks, using a heuristic algorithm utilizing Bayesian fitness function—matching the network structure to the probability distribution—as a metric.
- 2.
The conversion process allows for controlled modification of the acceptation level (PA) parameter. The research conducted in this direction shows that such development of the set of rules does not always help to improve the classification results. Is it therefore necessary to carry out studies evaluating the classification effectiveness of each developed set.
- 3.
The lack of the rule number is due to the fact, that during the development of the rule set, the rules are automatically numbered after applying the operation of grouping by the concept of the dependent variable and sorted within the concepts. In the result set the rule will be numbered 2. There is already a rule having the number in the set of rules presented in Fig. 5. To avoid ambiguity the authors deliberately omitted the numbering of the second generation rules.
References
Pearl, J.: Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan Kaufmann Publishers, San Mateo (1988)
Klopotek, M.: Intelligent Web Search Engine. EXIT, Warsaw (2001)
Duda, R.O., Hart, P.E., Nilsson, N.J.: Subjective bayesian methods for rule-based inference systems, Technical Note 124, SRI Project 4763 (1976)
Korver, M., Lucas, P.: Converting a rule-based expert system into a belief network. Med. Inform. 18(3), 219–241 (1993)
Shwe, M., Middleton, B., Heckerman, D., Henrion, M., Horvitz, E., Lehmann, H.: Probabilistic diagnosis using a reformulation of the INTERNIST-1/QMR knowledge base I. The probabilistic model and inference algorithms. Meth. Inf. Med. 30(4), 241–255 (1991)
Sniezynski, B.: Converting a naive bayes model into a set of rules. Intelligent Information Processing and Web Mining. Advances in Soft Computing, pp. 221–229. Springer, Berlin (2006)
Sniezynski, B.: Converting a naive bayes models with multi-valued domains into sets of rules. Database and EXpert Systems Applications, pp. 634–643. Springer, Berlin (2006)
Sniezynski, B.: Algorithm for converting bayesian network into set of rules: initial results. Technical Report TR-1/2007, AGH University of Science and Technology (2007)
Jensen, F.: Bayesian Networks and Decision Graphs. Springer, New York (2001)
Bachmann, P.: Die Analytische Zahlentheorie. Teubner, Leipzig (1984)
Grzymała-Busse, J.W., Hippe, Z.S., Mroczek, T.: Deriving belief networks and belief rules from data: a progress report. Transactions on Rough Sets VII. Lecture Notes in Computer Science, vol. 4400, pp. 53–69. Springer, Heidelberg (2007)
Grzymala-Busse, J.W., Hippe, Z.S., Kordek, A., Mroczek, T., Podraza, W.: Neonatal infection diagnosis using constructive induction in data mining. Rough Sets, Fuzzy Sets, Data Mining and Granular Computing. Lecture Notes in Computer Science (Lecture Notes in Artificial Intelligence), vol. 4482, pp. 289–296. Springer, Heidelberg (2007)
Grzymala-Busse, J.W., Hippe, Z.S., Mroczek, T., Paja, W., Bucinski, A., Strepikowska, A., Tutaj, A.: Brain stroke database—evaluation of glasgow outcome scale and the rankine scal. INFOBAZY (2008) Systems, Applications, Services. Academic Computer Centre in Gdansk—TASK 127–131(2008)
Grzymala-Busse, J.W., Hippe, Z.S., Mroczek, T., Roj, E., Skowronski, B.: Two Rough Set Approaches to Mining Hop Extraction Data. Information Science Reference, New York (2008)
Mroczek, T., Pancerz, K., Warchol, J.: Belief Networks in Classification of Laryngopathies Based on Speech Spectrum Analysis. Lecture Notes in Artificial Intelligence, pp. 222–231. Springer, Berlin (2012)
Gomuła, J., Paja, W., Pancerz, K., Mroczek, T., Wrzesień, M.: Experiments with hybridization and optimization of the rules knowledge base for classification of MMPI profiles. Advances in Data Mining. Lecture Notes in Computer Science, vol. 6870, pp. 121–133. Springer, Heidelberg (2011)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2016 Springer International Publishing Switzerland
About this paper
Cite this paper
Mroczek, T., Hippe, Z.S. (2016). Conversion of Belief Networks into Belief Rules: A New Approach. In: Burduk, R., Jackowski, K., Kurzyński, M., Woźniak, M., Żołnierek, A. (eds) Proceedings of the 9th International Conference on Computer Recognition Systems CORES 2015. Advances in Intelligent Systems and Computing, vol 403. Springer, Cham. https://doi.org/10.1007/978-3-319-26227-7_9
Download citation
DOI: https://doi.org/10.1007/978-3-319-26227-7_9
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-26225-3
Online ISBN: 978-3-319-26227-7
eBook Packages: EngineeringEngineering (R0)