Abstract
Ontology mapping is a common way to solve heterogeneous problem in ontologies and achieve the goal of ontology integration. In a distributed environment, it is necessary to find the mappings between ontologies before they are merged. There is a number of mapping approaches based on concept similarities which are widely used but with large amount of calculations exist. Aiming at mapping between ontologies and reducing computation complexity, we provide a new approach based on Classification with Word and CONtext Similarity (CWCONS) to find equivalence relation between concepts. The basic idea of CWCONS is to classify the nodes of ontology trees into two types, the classification nodes and the concept nodes, both of which rely on the tree structure of an ontology. We use Longest Common Substring (LCS) and Tversky's similarity model to evaluate similarities. Experimental results demonstrate that CWCONS can be more efficient and cost-effective with much less runtime over WCONS.