Logo eprints

Extending ontology queries with bayesian network reasoning

Bellandi, Andrea (2008) Extending ontology queries with bayesian network reasoning. Advisor: Turini, Prof. Franco. pp. 154. [IMT PhD Thesis]

[img]
Preview
Text
Bellandi_phdthesis.pdf - Published Version
Available under License Creative Commons Attribution No Derivatives.

Download (5MB) | Preview

Abstract

Today, ontologies are the standard for representing knowledge about concepts and relations among concepts concerning specific domains. In general, ontology languages are based on crisp logic and thus they can not handle incomplete or partial knowledge about application domain. However, uncertainty exists in domain modeling, ontology reasoning, and concept mapping. Our choice for dealing with uncertainty, is the Bayesian probability theory. The technique of representing an ontology by means of a bayesian network is used for having a new knowledge base enriched with uncertainty, over which making inference and probabilistic reasoning. Our method is composed of three steps. The first one is to compile the ontology into a bayesian network. We define the ontology compiling process for extracting the bayesian network structure directly from the schema of the knowledge base. The second one is to learn the initial probability distributions. We provide a computation process for learning the probability distributions, both prior and conditional, directly from the ontology instances, based on the Bayes theorem. The third one is to provide a bayesian query language for answering queries involving probabilities. Although evaluating bayesian networks is, in general, NP-hard, there is a class of networks that can efficiently be solved in time linear in the number of nodes. It is that of the polytree. On the basis of this bayesian network class, it is provided a bayesian query language for answering queries involving probabilities concerning both is - a ontology relations and object - property relations. It is based on recursive algorithms implementing top-down and bottom-up reasoning over polytrees networks.

Item Type: IMT PhD Thesis
Subjects: Q Science > QA Mathematics > QA75 Electronic computers. Computer science
PhD Course: Computer Science and Engineering
Identification Number: 10.6092/imtlucca/e-theses/8
Date Deposited: 05 Jul 2012 11:13
URI: http://e-theses.imtlucca.it/id/eprint/8

Actions (login required)

View Item View Item