Quantifying conditional probability tables in Bayesian networks: Bayesian regression for scenario-based encoding of elicited expert assessments on feral pig habitat

Alkhairy, Ibrahim; Low‐Choy, Samantha; Murray, Judith; Wang, Junhu; Pettitt, Anthony N.

doi:10.1080/02664763.2019.1697651

Cited by 4 publications

(5 citation statements)

References 65 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition, CPT helped to determine the future events of the risks relating to the end devices. The input data was generated and computed to determine the risk outcome (Alkhairy et al, 2020). The results of using the Tornado graphs demonstrated the risk impact and the likelihood of the end devices based on the available prior indicators to determine the level of the risk likelihood and extent of the risk impact.…”

Section: Discussionmentioning

confidence: 99%

Determination of the end device risk likelihood using the Bayesian network tools

Ncubukezi

2023

eccws

View full text Add to dashboard Cite

All institutions use end devices for information processing which includes sending and receiving on the network. This process helps them to improve their business production as well as perform daily activities at a faster rate. However, the increased usage of end devices by both employees and criminals raises concerns and exposes businesses to a range of cyber risks. End devices can sometimes be used as agents and weapons to expose internal business operations. The vulnerability of the end devices to cyber threats and attacks compromises business data, its safety, and security. This paper determines the risk likelihood of the end devices using the Bayesian network tools. To achieve this, the study illustrates the connections of the end device variables to simulate the risk likelihood and its impact. The analysis and interpretation of the simulation are performed using decision tree analysis (DTA), scenario analysis, and sensitivity analysis techniques (Tornado graphs, conditional probability tables (CPT), and value of information configuration (VOI)). The relationship of the variables is demonstrated on the AgenaRisk package. Results revealed variables that influence the risk probability and its impact. End device risks can be caused by insiders and cyber criminals. The risks associated with end devices are influenced by the level of security implementation on different levels. The impact of the cyber risks was also accounted for and the concluding remarks were also made.

show abstract

Section: Discussionmentioning

confidence: 99%

Determination of the end device risk likelihood using the Bayesian network tools

Ncubukezi

2023

eccws

View full text Add to dashboard Cite

show abstract

“…The approach can also be used similar to the weighted linear model approaches of Fenton et al (2007) and Alkhairy et al (2020), often a familiar format for many experts. In fact, InterBeta aims to combine the advantages of both interpolation approaches and weighted linear model approaches, while also allowing the number of parameters to be scaled to the desired level of fidelity to an expert's ideal CPT.…”

Section: Discussionmentioning

confidence: 99%

“…The approach is analogous to multivariable linear models as used in regression. Figure 4C shows a structure with nodes taking on labels common in such models, and recently Alkhairy et al (2017Alkhairy et al ( , 2020 have described a more explicit approach for using generalized linear models with a Beta link function for specifying the distribution over a binary child in much this fashion. Since structures like this are also at the heart of neural network nodes, we can represent such nodes by substituting in commonly used functions such as the sigmoid or softmax.…”

Section: Local Structuresmentioning

confidence: 99%

“…Perhaps the most well‐known and well‐supported is the Noisy‐OR (Kim & Pearl, 1983; Pearl, 1988). Other approaches include generalized Noisy functions (Srinivas, 1993), decision trees (Boutilier et al., 1996), algebraic decision diagrams (Chavira & Darwiche, 2007), weighted rank nodes (Fenton et al., 2007; Laitila & Virtanen, 2016; Noguchi et al., 2019) and discretized equations and interpolations (Alkhairy et al., 2020; Hassall et al., 2019; Mkrtchyan et al., 2016; Podofillini et al., 2014).…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A flexible method for parameterizing ranked nodes in Bayesian networks using Beta distributions

Mascaro

Woodberry

2022

Risk Analysis

View full text Add to dashboard Cite

When novice modelers first attempt to build a Bayesian network, they are often impressed with the intuitive graphical structures that capture their causal understanding. This favorable impression evaporates on proceeding to parameterization. Conditional probability tables (CPT) require parameters for often hundreds of very similar scenarios and specifying them in the absence of data can be overwhelming. The problem is even more severe when eliciting parameters from experts with limited time. Often, there is local structure with fewer parameters that better describes the relationship. Such structures include the Noisy OR, decision trees, and equations. These work well for modelers, but can be an issue for experts and particularly groups of experts. An alternative approach is to elicit only a few CPT rows and interpolate the remainder. This is a promising approach, as it can handle unknown structures and multiple experts, but existing techniques can be limited. Here, we present a flexible approach called InterBeta for performing CPT interpolation with ordered nodes. In the simplest case, just two CPT rows are needed, but this can be easily augmented with further information. The basic approach assumes input independence, but allows dependencies to be reintroduced as required, and can also be combined with other local structures such as decision trees or equations, leaving the interpolator to fill in the gaps. We explain the InterBeta method, describe its capabilities and limitations and how it compares to similar approaches and show how it can trade-off elicitation effort against faithfully representing expert understanding.

show abstract

“…The probability distribution of the unknown POS tag is calculated using a Bayesian network (BN) built on a sequence of POS tags. The BN is a probabilistic graphical model that makes explicit, through a directed acyclic graph, the interactions within a set of variables [7,8]. As for the second step, the paper considered two classifiers, in addition to the default Bayes Classifier, to summarize the above-mentioned distribution.…”

Section: Introductionmentioning

confidence: 99%

An auxiliary Part‐of‐Speech tagger for blog and microblog cyber‐slang

Golia

Zola

2022

Statistical Analysis

View full text Add to dashboard Cite

The increasing impact of Web 2.0 involves a growing usage of slang, abbreviations, and emphasized words, which limit the performance of traditional natural language processing models. The state-of-the-art Part-of-Speech (POS) taggers are often unable to assign a meaningful POS tag to all the words in a Web 2.0 text.To solve this limitation, we are proposing an auxiliary POS tagger that assigns the POS tag to a given token based on the information deriving from a sequence of preceding and following POS tags. The main advantage of the proposed auxiliary POS tagger is its ability to overcome the need of tokens' information since it only relies on the sequences of existing POS tags. This tagger is called auxiliary because it requires an initial POS tagging procedure that might be performed using online dictionaries (e.g., Wikidictionary) or other POS tagging algorithms. The auxiliary POS tagger relies on a Bayesian network that uses information about preceding and following POS tags. It was evaluated on the Brown Corpus, which is a general linguistics corpus, on the modern ARK dataset composed by Twitter messages, and on a corpus of manually labeled Web 2.0 data.

show abstract

Quantifying conditional probability tables in Bayesian networks: Bayesian regression for scenario-based encoding of elicited expert assessments on feral pig habitat

Cited by 4 publications

References 65 publications

Determination of the end device risk likelihood using the Bayesian network tools

Determination of the end device risk likelihood using the Bayesian network tools

A flexible method for parameterizing ranked nodes in Bayesian networks using Beta distributions

An auxiliary Part‐of‐Speech tagger for blog and microblog cyber‐slang

Contact Info

Product

Resources

About