Results 1 
5 of
5
Building probabilistic networks: where do the numbers come from?  a guide to the literature
 IEEE Transactions on Knowledge and Data Engineering
, 2000
"... Probabilistic networks are now fairly well established as practical representations of knowledge for reasoning under uncertainty, as demonstrated by an increasing number of successful applications in such domains as (medical) diagnosis and prognosis, planning, vision, ..."
Abstract

Cited by 32 (3 self)
 Add to MetaCart
Probabilistic networks are now fairly well established as practical representations of knowledge for reasoning under uncertainty, as demonstrated by an increasing number of successful applications in such domains as (medical) diagnosis and prognosis, planning, vision,
Sensitivity Analysis: an Aid for Beliefnetwork Quantification
, 2000
"... When building a Bayesian belief network, usually a large number of probabilities have to be assessed by experts in the domain of application. Experience shows that experts often are reluctant to assess all probabilities required, feeling that they are unable to give assessments with a high level of ..."
Abstract

Cited by 17 (7 self)
 Add to MetaCart
When building a Bayesian belief network, usually a large number of probabilities have to be assessed by experts in the domain of application. Experience shows that experts often are reluctant to assess all probabilities required, feeling that they are unable to give assessments with a high level of accuracy. We argue that the elicitation of probabilities from experts can be supported to a large extent by iteratively performing sensitivity analyses of the belief network in the making, starting with rough, initial assessments. Since it gives insight into which probabilities require a high level of accuracy and which do not, performing a sensitivity analysis allows for focusing further elicitation efforts. We propose an elicitation procedure in which, alternatingly, sensitivity analyses are performed and probability assessments are refined, until satisfactory behaviour of the belief network is obtained, until the costs of further elicitation outweigh the benefits of higher accur...
Properties of Sensitivity Analysis of Bayesian Belief Networks
 Proceedings of the Joint Session of the 6th Prague Symposium of Asymptotic Statistics and the 13th Prague Conference on Information Theory, Statistical Decision Functions and Random Processes, Union of Czech Mathematicians and Physicists
, 1999
"... The assessments obtained for the various conditional probabilities of a Bayesian belief network inevitably are inaccurate. The inaccuracies involved influence the reliability of the network's output. By subjecting the belief network to a sensitivity analysis with respect to its conditional prob ..."
Abstract

Cited by 13 (3 self)
 Add to MetaCart
The assessments obtained for the various conditional probabilities of a Bayesian belief network inevitably are inaccurate. The inaccuracies involved influence the reliability of the network's output. By subjecting the belief network to a sensitivity analysis with respect to its conditional probabilities, the reliability of the output can be investigated. Unfortunately, straightforward sensitivity analysis of a Bayesian belief network is highly timeconsuming. In this paper, we show that, by qualitative considerations, several analyses can be identified as being uninformative as the conditional probabilities under study cannot affect the network's output. In addition, we show that the analyses that are informative comply with simple mathematical functions; more specifically, we show that the network's output can be expressed as a quotient of two functions that are linear in a conditional probability under study. These properties allow for considerably reducing the computational burden of se...
Making sensitivity analysis computationally efficient
 Proceedings of the 16th Conference on Uncertainty in Artificial Intelligence (UAI
, 2000
"... To investigate the robustness of the output of a Bayesian network, a sensitivity analysis can be performed in which the relation between the output and each of the (probability) parameters of the network is established. This relation is given as a quotient of two linear functions in a parameter unde ..."
Abstract

Cited by 4 (0 self)
 Add to MetaCart
To investigate the robustness of the output of a Bayesian network, a sensitivity analysis can be performed in which the relation between the output and each of the (probability) parameters of the network is established. This relation is given as a quotient of two linear functions in a parameter under study. Current methods for computing the coefficients of these functions relies on a large number of probability propagations. In this paper, we present a method which only requires a single outward propagation in a junction tree for computing the coefficients associated with all the parameters, in addition to an inward propagation for processing evidence. Conversely, the method also only requires a single outward propagation for computing the coefficients associated with a single parameter and all possible outputs. We show that these results also hold for the analysis of the effects of joint variations of sets of parameters, known as nway sensitivity analysis.
Focused Quantification of a Belief Network Using Sensitivity Analysis
, 1999
"... The quantification of belief networks is known to be a laborious and difficult task, which hampers their application in practice. However, sensitivity analyses generally reveal that the influences of individual parameters on a network's performance differ considerably. This suggests that the qu ..."
Abstract
 Add to MetaCart
The quantification of belief networks is known to be a laborious and difficult task, which hampers their application in practice. However, sensitivity analyses generally reveal that the influences of individual parameters on a network's performance differ considerably. This suggests that the quantification effort can be focused on the most influential parameters, as for less influential parameters, rough estimates may suffice. The paper presents an empirical investigation of the viability of this approach, by comparing several belief network quantifications of different levels of informedness. It was established that refining a limited number of highly influential parameters in a poorlyinformed quantification may be sufficient to obtain satisfying network performance.