A framework for ontology integration based on genetic algorithm

Ontology integration is an important work when integrating information from heterogeneous ontologies into an ontology. The existing methods about ontology integration cannot effectively make full use of non-1-1 mappings, which are very common in the real world. Furthermore, these methods only stated that the concept-pairs with mappings should be integrated, but not gave the specific operations for it. Therefore, these methods cannot describe a complete framework for ontology integration. To this end, this paper proposes a framework for Ontology Integration based on Genetic Algorithm, called OI-GA. During the process of integrating ontologies, OI-GA firstly creates mappings between them based on similarity measures. Next, OI-GA finds out all the non-1-1 mappings from mappings, and provides an evolutionary method to extract 1-1 mappings from them. Finally, all the concepts belonging to different ontologies are integrated into a new knowledge base called integrated ontology. Experimental results indicate that OI-GA performs encouragingly well in the optimization of mapping set as well as in the integration of ontologies from the real world.