China
Africa
Explore chapters and articles related to this topic
Integrated hydrological risk analysis for hydropower projects
Published in Jean-Pierre Tournier, Tony Bennett, Johanne Bibeau, Sustainable and Safe Dams Around the World, 2019
T.H. Bakken, D. Barton, J. Charmasson
Bayesian networks are generic modeling tool used both for representing a correlation structure in a causal network and for decision analysis under uncertainty, by the means of a graphical representation of a joint probability distribution. They can be based on a combination of model results from ‘domain models’ (e.g. hydrological models) and on expert judgement using sparse data. Bayesian belief networks are increasingly being used to document and improve expert judgement in ecological modeling, decision support in the provision and demand of ecosystem services, and environmental and resource management, but have to a limited extent been used in the assessment of future water availability in regulated river basins with multiple water uses. Modelling platforms such as WEAP can explicitly represent a large number of water demands/requirements and hydrological risk elements. ‘Domain tools’ (e.g. WEAP) include, however, to only a limited extent risk elements that are most properly should be expressed qualita-tively, e.g. originating from expert judgments. There seems to be some scope for using Bayesian belief networks as a ‘meta modelling tool’ to integrate model simulation results on water availability with other risk sub-models. Such integrated models improve the decision-making process, usually not by pointing to an optimal decision, but by uncovering where in the causal chain investment monitoring provides the greatest information value for future decision-making.
Visual-Semantic Context Learning for Image Classification
Published in Spyrou Evaggelos, Iakovidis Dimitris, Mylonas Phivos, Semantic Multimedia Analysis and Processing, 2017
Qianni Zhang, Ebroul Izquierdo
There are many methods for learning both the structure and parameters of Bayesian networks from the given training data. Learning the network structure and parameters is in fact a search through the space of all possible links and parameters of the set of nodes ni, i = 1, 2, ..., t, where t is the total number of nodes. In the proposed SCL method, given a set of predefined classes and the training data, a visual-semantic context model is constructed by applying the K2 algorithm [239]. K2 is a greedy search technique. It starts from an empty network with random initial settings and creates a Bayesian network by iteratively adding a directed arc to a given node ni from the parent node whose addition most increases the K2 score of the resulting graph structure. The iterations terminate when no more possible additions could increase the K2 score. The evaluation metric for calculating the K2 score of a network structure is described as follows.
Adaptive Routing Provision by Using Bayesian Inference
Published in Jonathan Loo, Jaime Lloret Mauri, Jesús Hamilton Ortiz, Mobile Ad Hoc Networks, 2016
Ilias Kiourktsidis, Jonathan Loo, Grigorios Koulouras
A Bayesian network (BN) or belief network is a model. This model reflects the states of a system. The model can be anything; it can be a car, a human body, an ecosystem, a stock market, generally anything in the world. The objective is to create a BN that represents the communication between two peers in a MANET. Practically, a Bayesian network is a probabilistic graphical model that represents the joint probability distribution for a set of variables via a directed acyclic graph. The variables are some characteristics of the system that the Bayesian network models. For a car, the variables can be the car engine, the tires, the speed, the car’s age, etc. A Bayesian network that models a car represents the probabilistic relations between the variables of the car we just mentioned. Given the state (condition) of the car engine and tire variables, the network can compute the probability distributions of the speed of the car and the age of the car variables or vice versa. Typically some states will tend to occur more frequently when other states are present. Thus, if you are sick, the chances of a runny nose are higher. If it is cloudy, the chances of rain are higher, and so on.
Inference of HDVs real-time locations in mixed autonomous traffic flow scenario
Published in Transportmetrica B: Transport Dynamics, 2022
Hongsheng Qi, Peng Chen, Yuyan Ying
The relationship shown in Figure 4(b) can be depicted in various ways, for instance, using neural network, nonlinear regression, etc. In this study, we employ the BN to construct such relationship, as BN can encode the influence among variables in a flexible way. Once the relationship between vehicle state variables is constructed, the inference can be made easily. We name the relationship which employs the BN as CF-BN. Besides, there are two types of Bayesian network: static Bayesian network and dynamic Bayesian network. The dynamic one explicitly takes the temporal evolution into account. In our case, the nodes are the vehicles’ states (acceleration, speed, or longitudinal displacement, etc.) at certain moments. Because these variables are dynamic, the static Bayesian network structure is adopted by iteratively applying the CF-BN. In this way, the computational efficiency can be guaranteed.
A comparison between the Bayesian network model and the logistic regression model in prevention of the defects on ceramic tiles
Published in Journal of Experimental & Theoretical Artificial Intelligence, 2022
Volkan Sevinç, Meryem Merve Kırca
Bayesian network is a type of machine learning, which is a branch of artificial intelligence. The main idea of Bayesian networks is to reflect the conditional probabilistic relations of the variables to a graphical model. Bayesian networks, which were first introduced by Pearl (1985), have two main parts as the graphical and the probabilistic structures. The graphical structure of the Bayesian networks is based on a visual architecture called directed acyclic graph (DAG). The structure of a DAG consists of nodes and directed arrows (edges), which connect the nodes. The nodes represent the variables, and the arrows indicate the conditional probabilistic relations among the nodes. Estimations of the conditional probabilistic relations in Bayesian networks are obtained by creating tables called conditional probability tables (CPT), which form the probabilistic structure of the Bayesian networks.
Beyond Purchase Intentions: Mining Behavioral Intentions of Social-Network Users
Published in International Journal of Human–Computer Interaction, 2022
Dynamic Bayesian networks (Murphy, 2002) are an extension of static Bayesian networks that allow for a dynamic representation of temporal nodes and edges. The use of dynamic Bayesian networks, a relatively new modeling technique, in the existing literature is less common than the use of Bayesian networks. The main areas to which dynamic Bayesian networks have been applied thus far are safety monitoring and reliability analysis, disaster prediction, risk assessment, and biology. Dynamic Bayesian networks were used to model potential threats posed by dynamic network vulnerabilities (Frigault et al., 2008) and for predicting insider threats (Axelrad et al., 2013); reliability evaluation and safety decision support (Amin et al., 2018; Li et al., 2017); modeling gene expression data (Dojer et al., 2006; Murphy & Mian, 1999); predicting disasters, such as car accidents, banking crisis, and wildfires (Dabrowski et al., 2016; Khakzad, 2019; Sun & Sun, 2015). Some works also used dynamic Bayesian networks for user modeling. Examples include human driving behavior (Kumagai & Akamatsu, 2006), students’ learning styles (Käser et al., 2017), and user stress and anxiety levels (Liao et al., 2005).