Using sensitivity analysis for efficient quantification of a belief network

Sensitivity analysis is a method to investigate the effects of varying a model's parameters on its predictions. It was recently suggested as a suitable means to facilitate quantifying the joint probability distribution of a Bayesian belief network. This article presents practical experience with performing sensitivity analyses on a belief network in the field of medical prognosis and treatment planning. Three network quantifications with different levels of informedness were constructed. Two poorly-informed quantifications were improved by replacing the most influential parameters with the corresponding parameter estimates from the well-informed network quantification; these influential parameters were found by performing one-way sensitivity analyses. Subsequently, the results of the replacements were investigated by comparing network predictions. It was found that it may be sufficient to gather a limited number of highly-informed network parameters to obtain a satisfying network quantification. It is therefore concluded that sensitivity analysis can be used to improve the efficiency of quantifying a belief network.

[1]  Michael P. Wellman.  Fundamental Concepts of Qualitative Probabilistic Networks , 1990, Artif. Intell..

[2]  Niels Peek,et al.  Explicit temporal models for decision-theoretic planning of clinical management , 1999, Artif. Intell. Medicine.

[3]  A. Tversky,et al.  Judgment under Uncertainty , 1982 .

[4]  L. C. van der Gaag,et al.  Practicable sensitivity analysis of Bayesian belief networks , 1998 .

[5]  大西 仁,et al.  Pearl, J. (1988, second printing 1991). Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference. Morgan-Kaufmann. , 1994 .

[6]  J D Habbema,et al.  Analysing clinical decision analyses. , 1990, Statistics in medicine.

[7]  Linda C. van der Gaag,et al.  Sensitivity analysis: an aid for belief-network quantification , 2000, The Knowledge Engineering Review.

[8]  Giuseppe Carenini,et al.  Dealing With the Expert Inconsistencies: The Sooner the Better , 1995 .

[9]  Niels Peek,et al.  Developing a Decision-Theoretic Network for a Congenital Heart Disease , 1997, AIME.

[10]  D. Heckerman,et al.  Toward Normative Expert Systems: Part I The Pathfinder Project , 1992, Methods of Information in Medicine.

[11]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[12]  J. Moller,et al.  Late results (30 to 35 years) after operative closure of isolated ventricular septal defect from 1954 to 1960. , 1991, The American journal of cardiology.

[13]  Campbell Cf,et al.  Ventricular septal defects--what's new? , 1970 .

[14]  W. Edwards,et al.  Decision Analysis and Behavioral Research , 1986 .

[15]  Marek J. Druzdzel,et al.  Elicitation of Probabilities for Belief Networks: Combining Qualitative and Quantitative Information , 1995, UAI.

[16]  A. Moss,et al.  Heart disease in infants, children, and adolescents , 1977 .

[17]  Steen Andreassen,et al.  MUNIN - A Causal Probabilistic Network for Interpretation of Electromyographic Findings , 1987, IJCAI.

[18]  K Sarkar,et al.  Spontaneous closure of ventricular septal defect. , 1974, Journal of the Indian Medical Association.

[19]  Gregory M. Provan,et al.  The Sensitivity of Belief Networks to Imprecise Probabilities: An Experimental Investigation , 1996, Artif. Intell..

[20]  Max Henrion,et al.  Uncertainty: A Guide to Dealing with Uncertainty in Quantitative Risk and Policy Analysis , 1990 .