Explore chapters and articles related to this topic
Driver Behavior
Published in Motoyuki Akamatsu, Handbook of Automotive Human Factors, 2019
Once the structure of the BN is determined and some of the observed variables or probability distributions are obtained, the probability distributions of other variables can be inferred (Murphy, 2002). To calculate exact solutions, the junction tree algorithm is used to convert a graph into a tree structure, and the belief propagation algorithm is used to obtain marginal distribution. It is important that Xi depends only on Pa (Xi) and does not depend on any of the variables. This fact makes it possible to reduce the amount of calculation required for probabilistic inference and to study the conditional independence of and dependency relationships between variables.
Signature Generation Algorithms for Polymorphic Worms
Published in Mohssen Mohammed, Al-Sakib Khan Pathan, Automatic Defense Against Zero-day Polymorphic Worms in Communication Networks, 2016
Mohssen Mohammed, Al-Sakib Khan Pathan
For multiply connected graphs, the standard exact inference algorithms are based on the notion of a junction tree [82]. The junction tree algorithm is one of the most widely used algorithms, the basic idea of this algorithm is to group variables to convert the multiply connected graph into a singly connected undirected graph (tree) over sets of variables and do inference in this tree [36].
Uneven spatial distribution of fatigue cracks on steel box-girder bridges: a data-driven approach based on Bayesian networks
Published in Structure and Infrastructure Engineering, 2021
Jiahui Chen, Dongyu Zhang, Wensong Zhou, Zhicheng Chen, Hui Li
After determining the CPTs, probabilistic inference can be conducted using junction tree algorithm first proposed by Lauritzen and Spiegelhalter (1988). In this section, the test dataset is used to verify the BN. For all the girder-elements in the test dataset, the probability of each type of crack can be calculated by the established BN. For example, if we want to know the probability of a Type I crack on an girder-element with variable of critical variable being (located in lane 1), (located at position 1), (8 mm-thick diaphragm), (2500 mm long), (14 mm-thick deck), (low-temperature zone) respectively; then, the probability of occurrence of type I crack can be calculated as:
Modelling the spatial distribution of heavy vehicle loads on long-span bridges based on undirected graphical model
Published in Structure and Infrastructure Engineering, 2019
Zhicheng Chen, Yuequan Bao, Jiahui Chen, Hui Li
When the number of nodes in the vehicle-location UGM is large, the joint PMF cannot be directly calculated via Equation (4) because it involves the calculation of the partition function which requires specifying all possible combinations of realizations of the random vector given in Equation (2). An alternative approach is using the junction tree algorithm, which can provide an exact inference procedure for arbitrary graphical models and highly efficient if the treewidth is small (Murphy, 2012). The so-called treewidth is defined as the number of nodes contained in the largest clique of the junction tree. The junction tree built from the vehicle-location UGM has a small treewidth being equal to where is the number of lanes.