Henrion, M. (1990). Towards efficient probabilistic diagnosis in multiply connected networks. In Oliver, R. and Smith, J., editors, Influence Diagrams, Belief Networks, and Decision Analysis, pages 385--409. John Wiley and Sons, Chichester.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....First, both the qualitative and quantitative causalities of the interrelated hypotheses can be modeled in form of a simple graph. Second, the knowledge and beliefs of human experts can be structured in a coherent manner so that the consistency and completeness of the knowledge in a network [2] can be guaranteed. Applications of Bayesian network representation for the development of knowledgebased systems can be found in several domains; for example, clinical decision making [3] medical diagnosis [2,4,5] and the assessment of the design methodologies for an AI based augmentative ....

....coherent manner so that the consistency and completeness of the knowledge in a network [2] can be guaranteed. Applications of Bayesian network representation for the development of knowledgebased systems can be found in several domains; for example, clinical decision making [3] medical diagnosis [2,4,5], and the assessment of the design methodologies for an AI based augmentative communication device [6,7] Although Bayesian network representation has been widely used in various disciplines, most of the current applications of Bayesian networks have concentrated on deriving the exact belief value ....

[Article contains additional citation context not shown here]

M. Henrion, "Towards Efficient Probabilistic Diagnosis in Multiply Connected Belief Networks," in Influence Diagrams, Belief Nets and Decision Analysis, editors: R. Oliver and J. Smith, John Wiley & Sons Ltd., pp. 385--409, 1990.

.... this representation technique and the semantics of uncertain knowledge [12,14,19] In parallel to the progress on the theoretical foundation of knowledge representation, advances have also been made in the algorithmic development of efficient inference schemes for use in a Bayesian belief network [5,11,16,17,21,22]. Recently the feasibility and the utility of applying Bayesian belief network to develop diagnostic systems of various domains have been demonstrated. These encouraging news have resulted in a significant increase in the application of Bayesian belief network modeling across various domains ....

Henrion M., 1990. Towards Efficient Probabilistic Diagnosis in Multiply Connected Belief Networks, in R.M. Oliver and J.Q. Smith eds, Influence Diagrams, Belief Nets and Decision Analysis, pp. 385-409, John Wiley & Sons Ltd., New York.

....hypotheses concerning the variables b; f; g; h; i and j in Fig. 1, given the observation S e = AC: 4 BF G H I J BF G H IJ B FG H I J Applications of this type of query are of great use and importance in several domains such as medical diagnosis [9] and clinical decision making [10] An efficient recurrence local computation algorithm has been developed [11] for the derivation of the partial ordering of composite hypotheses. The computational complexity of the algorithm is in the order of O(lkm) for a singly connected network; where l is the ....

M. Henrion, "Towards Efficient Probabilistic Diagnosis in Multiply Connected Belief Networks," in Influence Diagrams, Belief Nets and Decision Analysis, editors: R. Oliver and J. Smith, John Wiley & Sons Ltd., pp. 385-409, 1990.

.... Bayesian network is NP hard [2] Various research efforts have been focused on the restricted cases such as reasoning only the likelihoods of simple hypotheses that each of which consists of only one random variable [1,3] or on a special type of network configurations such as BN2 networks [4,5]. We have found in our previous research [6,7,8] that another possible direction of tackling the problem is to relax the constraints about the restricted type of networks and the restricted type of hypotheses, but to pursue probabilistic inference aiming at a partial solution, rather than a ....

M. Henrion, "Towards Efficient Probabilistic Diagnosis in Multiply Connected Belief Networks," in Influence Diagrams, Belief Nets and Decision Analysis, editors: R. Oliver and J. Smith, John Wiley & Sons Ltd., pp. 385--409, 1990.

....[Jen95] 2.6.3 Sensitivity Analysis How sensitive are answers to model probabilities: Kor90] HS93] Las93] CNKE93] NK91] NA91] Pro91] CS95] Derivatives: Bun95a] 2.6. 4 K most probable cases Compute the K most probable configurations, rather than just the single most probable one: Hen90] SG94] These find K highly probable plans but no necessarily the K best: KNL 93] 2.6.5 Generating Explanations Explaining an inference: Shi93] LD93] 2.6.6 Other Obtaining a simple description of the optimal policy in an Influence diagram: LS93] 3 Learning Acquiring Models Survey: ....

.... [LL94] SE94] NB94] Pro94] PFH94] Had94] XPB92] BGHK94] Bre92] NH95b, NH95a] HHNK95, NHH95] GC93, GC90] DS94] Pro93a] GK95] Fusing Multiple Networks: MA93] Sha91] 5 Applications Surveys: HMW95] and CACM March 1995 issue) NO93] Medical: BBQS92] Coo84] Hen90] SEH90] dBP90] Spi87] Spi90] Hec89] DG93b] BBS91] KRW 95] Pro93a] Monitoring (non medical) HB95] Tracking: KC90] Fun88] HKM 94] High Risk Analysis: OY90] Nuclear: PH90] Troubleshooting: HBR94] BH95] Vision: JCJ93] JCN92] Classification: MC93] ....

Max Henrion. Towards efficient probabilistic diagnosis in multiply connected belief networks. In R. M. Oliver and J. Q. Smith, editors, Influence Diagrams, Belief Nets, and Decision Analysis, chapter 17, pages 385--409. John Wiley & Sons, 1990.

....IB MAPs show a great performance improvement. The above cited papers [8, 21] as well as this one, are essentially deterministic approximation algorithms. Comparison with stochastic approximation algorithms should also be interesting. Stochastic simulation to approximate marginal probabilities [15] is one such stochastic algorithm. We do not have a ready performance comparison, and the method does not seem immediately applicable to this work. Other stochastic approximation algorithms find the MAP. For example, in [11] simulated annealing is used. It is not clear, however, how one might use ....

Max Henrion. Towards efficient probabilistic diagnosis in multiply connected belief networks. In J. Q. Smith and R. Oliver, editors, Influence Diagrams for Decision Analysis, Inference and Prediction. John Wiley and Sons, in preparation.

.... and Reggia s application of minimal covering sets of diseases to explain observed findings [114] and Henrion s TopN algorithm, which imposes a heuristic search on top of a branch and bound algorithm to prune inadmissible paths from the search tree, and to identify the n most probable diagnoses [67]. BN2Os are BN2s with noisy OR gates and leaks. These additions trade elegance for realism by recognizing that some observations may arise spontaneously without an explicitly modeled cause [69] Methods for solving BN2O s include a variant of TopN [69] a general approximation method called random ....

M. Henrion. Towards efficient probabilistic diagnosis in multiply connected networks. In R. M. Oliver and J. Q. Smith, editors, Influence Diagrams, Belief Nets, and Decision Analysis, chapter 17, pages 385--407. Wiley: Chichester, 1989.

..... 4 Lack of a Gold Standard Posterior Distribution Ideally, to examine the convergence properties of the simulation that we are using, we would like to know the posterior distribution implied by the QMR DT model that is, a gold standard distribution. It is possible that Henrion s TopN algorithm [11] would be able to produce tight bounds on the posterior probabilities of diseases for large cases. In addition, the recursive decomposition algorithm [12] appears promising as an exact method to calculate the posterior probabilities of diseases in an acceptable amount of time in some cases. As ....

Henrion M. Towards efficient probabilistic diagnosis in multiply connected networks. In: Oliver RM, Smith JQ, eds. Influence Diagrams, Belief Nets and Decision Analysis. Chichester: Wiley, 1990: 385-407. Middleton B, Shwe M, Heckerman D,et al. 31

No context found.

Henrion, M. (1990). Towards efficient probabilistic diagnosis in multiply connected networks. In Oliver, R. and Smith, J., editors, Influence Diagrams, Belief Networks, and Decision Analysis, pages 385--409. John Wiley and Sons, Chichester.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC