摘要
貝氏網路(Bayesian network)在決策分析的工具中，被應用於故障設施的判斷、醫療診斷、礦物探勘、以及影像辨識等各種領域。然而，在建構貝氏網路時，常出現專家意見與歷史資料不一致的情況。面對貝氏網路中一致性修正的部分，學者提出的方式不盡相同。如何在不同領域與問題特性下，選擇有效的一致性修正方式，成為一個重要的問題。本研究回顧學者提出的不同方法。並應用台電故障設備之歷史資料，以及Chien et al. (1999)根據歷史資料與專家知識所建構之貝氏網路，比較不同學者提出之一致性修正方法，做實證研究。此外，本研究建構一個新的方法，處理一致性修正的問題，並與學者所提出之方法做比較。本研究之目的為應用台電歷史資料與已建構的貝氏網路，驗證各種方法所得之結果，與歷史資料的相關性，並經由分析的結果，期望在此後面對一致性問題時，提出一個最為有效的方法，或者面對不同領域時，建議一個較為有效適用的修正方式。

關鍵字：貝氏網路、一致性修正、決策分析、知識管理

Abstract
The Bayesian Network is a probabilistic graphical model in which a problem is structured as a set of parameters and probabilistic relationships among them. That is, a Bayesian Network for the domain represents a joint probability distribution over those variables. It has been used to effectively incorporate the expert knowledge and the historical data for updating the prior belief in the light of new evidence in many fields. However, few researchers have studied the comparison of different methods for resolving the inconsistency between priors and inference rules of a Bayesian Network. This study aimed to compare the existing methods that modified the inconsistency between priors and inference rules and use empirical data for demonstration. In particular, Chien et al. constructed a Bayesian Network on the basis of expert knowledge and historical data for fault diagnosis on distribution feeder in Taiwan. We also developed a new method to modify the inconsistency between priors and inference rules of a Bayesian Network for comparison.

Keywords: Bayesian network, Inconsistency modification, Decision analysis, Knowledge management

Reference
Buchanan, B. G. and Shortliffe, E. H., Rule-Based expert Systems: The MYCIN Experiments of the Stanford Heuristic Programming Project, Addison-Wesley Publishing Company, Menlo Park, California, 1984.
Buntine, W., "A Guide to the Literature on Learning Probabilistic Networks from Data", IEEE Transactions on Knowledge and Data Engineering, vol.8, no.2, pp. 195-210, 1996.
Charniak, E., "Bayesian Networks without tears", AI Magazine, vol.12, no.4, pp. 50-63, 1991.
Chien, C. F., Chen S. L., Lai H. L., and Lin Y. S., "Using Bayesian Network for Fault Location on Distribution Feeder of Electric Power System", paper submitted 1999.
Clemen, R. T., Make Hard Decisions: An Introduction to Decision Analysis, Duxbury Press, California, 1996.
Duda, R. O., Gaschning, J., and Hart, P. E., "Model design in the prospector consultant system for mineral exploration," in: Michie, D. (Eds.), Expert Systems in the Microelectronic Age, Edinburgh University Press, Edinburgh, pp. 153-167, 1979.
Duda, R. O., Hart, P. E., and Nilsson, N. J., "Subjective Bayesian methods for rule-based inference systems," Proceeding 1976 National Computer Conference, vol.45, pp. 1075-1082, 1976.
Duda, R. O., and Shortliffe, E. H., "Expert system research," Science, vol. 220, pp. 261-268, 1983.
French, S., Decision Theory: An Introduction to the Mathematics of Rationality, Ellis Horwood Limited, London, England, 1993.
Henrion, M., Breese, J. S., and Horvitz, E. J., "Decision analysis and Expert systems", AI Magazine, vol.12, no.4, pp. 64-91, 1991.
Kirkwood, C. W., Strategic Decision Making: Multiobjective Decision Analysis with Spreadsheets, Duxbury, Belmont, California, 1997.
P., Poza M., Yurramendi Y., Murga R. H., and Kuijpers C. M. H., "Structure Learning of Bayesian Networks by Genetic Algorithms: A performance Analysis of Control Parameters", IEEE Transactions on Pattern analysis and Machine Intelligence, vol.18, no.9, pp. 912-926, 1996.
Laskey, K. B., "Sensitivity Analysis for Probability Assessments in Bayesian Networks", IEEE Transaction on systems, mans, and cybernetics, vol.25, no.6, pp. 901-909, 1995.
Lauritzen, S. L., and Spiegelhalter, D. J., "Local Computations with Probabilities on Graphical Structures and their Application to Expert system", Royal Statistics Society, no.2, pp. 157-224, 1988.
Pearl, T., Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference, Morgan Kaufamann Publishers, INC., San Mateo, California, 1988.
Pereira, C. A. D. B., and Barlow, R. E., "Medical Diagnosis Using Influence Diagrams", NETWORKS, vol.20, pp. 565-577, 1990.
Poole, D., "Probabilistic conflicts in a search algorithm for estimating posterior probabilities in Bayesian networks", Artificial Intelligence, vol. 88, pp. 69-100, 1996.
Shachter, R. D., "An Ordered Examination of Influence Diagrams", Networks, vol.20, pp. 535-563, 1990.
Shortliffe, E. H., MYCIN: A ruled-based Computer Progrm for Advising Physicians Regarding Antimicrobial Therapy Selection, Stanford Artificial Intelligence Laboratory Memo AIM-251, Stanford, California, 1974.
Shortliffe, E. H., and Buchanan, B.G., "A Model of Inexact Reasoning in Medicine", Mathematical Biosciences, vol. 23, pp. 351-379, 1975 (cited in Duda et al., 1976).
Sovarong, L., and Costas, J. S., "A General Equipment Diagnostic System and its Application on Photolithographic Sequences", IEEE Transactions on Semiconductor Manufacturing, vol.10, no.3, pp. 329-343, 1997.
Wu, Y. C., and Gustafson, D. H., "Removing the assumption of conditional independence from Bayesian decision models by using artificial neural networks", Artificial Intelligence in Medicine, vol. 6, pp. 437-454, 1994.