An ILP Approach to Ontology Refinement for the Semantic Web

A poster presentation written by Francesca Alessandra Lisi. It was presented at the ESWC2007. It is about Semantic Web Mining, Inductive Logic Programming, Ontology Management, Ontology Learning and Semantic Web Rules

[edit] Abstract

Ontology Refinement is a phase in the Ontology Learning process that aims at the adaptation of an existing ontology to a specific domain or the needs of a particular user. In this paper we consider the problem of refining a known concept (reference concept) belonging to a given taxonomic ontology in the light of new knowledge coming from an external data source. Ontologies for the Semantic Web are represented with mark-up languages based on Description Logics (DLs) and intended to be integrated with relational data to form rules based on some extension of Horn Clausal Logic (HCL) in the style of AL-log. We present a novel approach to the Ontology Refinement problem in hand which relies on the methodological apparatus of Inductive Logic Programming (ILP) and adopts the representation framework of AL-log. The approach borrows ideas from Concept Formation and Frequent Pattern Discovery, and assumes the existing ontology and the external data source are represented with the DL ALC and the function-free HCL Datalog respectively. Output concepts are represented as AL-log rules and are formed out of frequent patterns associating the reference concept and other concepts already occurring in the input ontology.

This data has been imported from the ESWC2007 RDF