We propose a method to align di erent ontologies in similar domains and then de ne correspondence between concepts in two di erent ontologies using the SKOS model. Introduction. Recently ontologies are created to provide knowledge representation. They use common representation languages such as OWL, but there are many heterogeneous ontologies [1{3]. In this paper we rst propose a lexical and structural analysis and compute the concept similarity as a combination of attributes, second use the SKOS model to de ne correspondence between concepts[4]. Ontology Alignment Framework.To perform the matching between concepts in di erent ontologies, we focus both on syntactical and text in entity descriptions and also their semantic structure in the ontology representations. This process, illustrated in the block diagram shown in Figure 1, is divided into two main sub-tasks: Alignment and SKOS translation. The inputs are two ontologies and result of the process is an SKOS-based ontology that contains automatically de ned associations.The alignment task analyses lexical and structural attributes of ontologies to automatically produce associations between concepts. The relation is de ned: R(A; B) =< A; B; Relation; S(A; B) > where A and B are ontology concepts, Relation describe semantic relations between these concepts which have ve types: equal beIncluded, include, disjoint, related, and S(A; B) is similarity measure for two concepts based on their structure and lexical analysis.
De ning SKOS-based Associations. After identifying possible relations between concepts, they are imported based on the SKOS model. This will provide an interconnection between two ontologies based on standard set of properties dened in the SKOS model. The SKOS mapping properties include skos:closeMatch, skos:exactMatch, skos:broadMatch, skos:narrowMatch and skos:relatedMatch. The properties maintain a mapping between SKOS concepts adapted from schemes. The relations in concept pairs de ned in the previous section are based on synset relations in WordNet. They are obtained according to accessing the extended synset collection for each representative word that describes entities and calculating structural similarity We will map between synset and SKOS relations. By applying these mappings, the nal product of the ontology integration process will include assertion axioms in which the related concepts from di erent ontologies are linked to each other based on SKOS relations. The integrated ontology will be a collection of concepts and properties from both ontologies and will also include the SKOS association properties. Figure 2 illustrates a part of the SKOS relations and concept alignment between two ontologies from the dataset (a complete set of our evaluation results using OAEI2008 dataset can be accessed from: http://tinyurl.com/38veolh).
Acknowledgement.The work is partially supported by the m:Ciudad project funded by the European 7th Framework Programme, contract number: 215007. Xiulei Liu's and Jianxin Liao's research is supported by Chinese National Science Fund for Distinguished Young Scholars (No. 60525110) and Chinese National 973 Program (No. 2007CB307100, 2007CB307103).