A Statistical Amalgamation Approach for Ontologies
As ontology is subjective and varies in different domains, the amount of ontologies turns out to be huge but with poor compatibility. Mainstream method for ontology integration is mostly achieved by establishing mappings between ontologies. In this essay, the author put forward another way of ontology merging. After statistic machine learning on concept relations, the frequency of different ontologies appeared in concept relations reveals certainty factor and help to build a large-scale concept relations network including the statistic information and domain categories, so that the conceptions conveyed by different ontologies can be fused together and the merging concept space turns to be relatively objective.