Critical Remarks on Bayesian Network Libraries
نویسنده
چکیده
Designing a large Bayesian network (BN) has been regarded as a diicult process. It has been suggested that BN libraries can be used to facilitate the construction of a large BN. That is, a large BN can be deened in terms of smaller BNs stored in a library. In this paper, we point out that it may be possible to combine the conditional independen-cies deened by the smaller BNs, but not the smaller BNs themselves.
منابع مشابه
Critical Remarks on the Maximal Prime Decomposition of Bayesian Networks
We present a critical analysis of the maximal prime decomposition of Bayesian networks (BNs). Our analysis suggests that it may be more useful to transform a BN into a hierarchical Markov network.
متن کاملPollock on Probability in Epistemology
John Pollock did a lot of interesting and important work on the metaphysics and epistemology of probability over several decades. In Thinking About Acting [15], we find many fascinating and thought provoking ideas and arguments (both old and new) about probability. Owing to limitations of space, I will be confining my remarks to a handful of issues addressed in [15] pertaining to probability, l...
متن کاملDynamic Safety Analysis CNG Stations Using Fault Tree Approach and Bayesian Network
Introduction: The safety of CNG stations is important because of their location in urban areas, as well as to prevent accidents and to protect the safety of personnel, property, and environment. An event occurrence analysis with probability updating is the key to dynamic safety analysis. Methods and materials: In this study, the Failure Modes and Effects Analysis (FMEA) technique was used to d...
متن کاملRecommending Library Methods: An Evaluation of Bayesian Network Classifiers
Programming tasks are often mirrored inside an organisation, across a community or within a specific domain. We propose that final source codes can be mined, that knowledge and insight can be automatically obtained and that this knowledge can be reused for the benefit of future developments. We focus on reusable software libraries; we wish to learn information about how such libraries are used ...
متن کاملDynamic Bayesian Information Measures
This paper introduces measures of information for Bayesian analysis when the support of data distribution is truncated progressively. The focus is on the lifetime distributions where the support is truncated at the current age t>=0. Notions of uncertainty and information are presented and operationalized by Shannon entropy, Kullback-Leibler information, and mutual information. Dynamic updatings...
متن کامل