Abstract
dc:description.abstractCausal Bayesian Networks are a widely recognised tool for modelling the uncer- tainty of a wide range of processes, particularly when the nature of how different factors influence each other. The practice of utilising Causal Bayesian Network is now becoming a growing trend for business that want to fully understand the demands im- posed on them, and how best to adapt their business in order to be successful. When designing and building a Causal Bayesian Network, it is often necessary to consult with domain experts for information about the shape of the model but also the defini- tion of how the causal factors influence others. The definition of these influences can require the specification of a large volume of probability distributions, even if a lot of evidential data is available for analysis. Whilst the definition of the structure of the model can be a relatively simple task for a domain expert, providing the probability distributions is a much more difficult task. In this thesis I discuss a method whereby, given a model structure, a domain expert can provide simple descriptive meta-data so that a hypothetical probability distribution can be generated for the discrete model variables.
Degree
thesis:*- Name dc:type.qualificationname
- M.Sc by research
- Level dc:type.qualificationlevel
- masters
- Grantor dc:publisher.institution
- University of York
- Year dc:date.issued
- 2017
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Balmer, Mark
- Advisor dc:contributor.advisor
-
- Cussens, James
Rights
- Language dc:language.dcterms_iso639_2
- eng