ABSTRACT
While building ontology, viewpoints of developers can be different. To support sharing of IPM ontologies developed by various agricultural experts, ontology merging is required. Creating a new ontology from two source ontologies is known as ontology merging. IPM ontologies developed by distinct agricultural experts can be integrated together to enrich knowledge base of IPM practices for a specific crop. The research work introduces here IPMOntoShare, a system for mapping and merging two IPM ontologies. Comparison of similarity of concepts at each level of ontology is discussed in detail.
In this chapter, special notations for ontology merging, known as Ontology Merging (OM) notations, are presented. Ontology merging algorithm is also presented where merging takes place automatically without user intervention. A list of labels to be used in ontology, which identifies the description of the concepts and relations among them, is presented. In OM algorithm, similarity of concepts from two ontologies is measured. Similar concepts are added once to a new third ontology and all differing concepts from both ontologies are added as a new concept in third ontology. Relation similarity is also considered. For concepts which are synonyms of each other, COM algorithm of OM is used. Removal of nested relations is also discussed.
An ontology merging method based on WordNet is proposed. It is specifically developed for merging heterogeneous domain ontologies. In this work, sets of concepts are prepared. Equality of concepts is measured after that. WordNet Synset number is used for this purpose. Values of sets of concepts are also measured using Jaccord coefficient and subclass relationships using most-specific-parent method. Hypernym and hyponym relations are checked for using PTR field in WordNet. Hierarchy is reconstructed based on the measurements.
