Results from experiments in ontology learning including evaluation and recommendation
Abstract
This document describes the Ontology Learning experiments we performed in order to recommend a set of ontology learning techniques to enhance the ontology engineering process in the fisheries domain in place at FAO. Experiments have been conceived in the wider context of the WorkPackage 7 of the NeOn Project. The main contribution of this document to WP7 is a set of recommendations and best practices to exploit semi-automatic technique to acquire knowledge either from domain specific documents and existing ontologies. The basic criterion for success indication adopted is the reduction of the development time required for the Ontology Engineering process actually in place at FAO. After a brief overview of WP7, the document addresses the following issues: to describe the state of the art in ontology learning and; to set up an evaluation case study in the fisheries domain; to identify suitable techniques which can be profitably applied to fit the user’s requirements; to evaluate such techniques in the use case scenario, and, finally, to provide a set of recommendations indicating the most reliable techniques to be included in the ontology engineering lifecycle.