Automatic Ontology Mapping for Agent Communication in an e-Commerce Environment

Internet-based e-commerce provides a high level of flexibility and openness though presenting many drawbacks due to the heterogeneity of the exchanged information. Ontologies are a key technology to solve many of the problems of e-commerce, in fact many companies use ontologies as a method of exchanging meaning between different agents. As ontology usage becomes more prevalent, the need for ontology reconciliation increases. In fact, ontology mapping methods can contribute to solve the problem of knowledge communication and interchange. In this paper we present an automatic method for ontology mapping. The method is made up of two phases: a lexical-semantic analysis based on the WordNet thesaurus and a structural analysis based on a matching algorithm that finds semantic mappings between two ontologies expressed in Attributive Language with Number description (ALN) Description logic. The mapped ontologies describe the same conceptualization through a set of rules that join related concepts. We deployed the proposed approach in a prototype system that currently is employed for large scale experiments. A simple experiment with a case study domain has shown a good correspondence with human mapping manually conducted and the system provided results.