Vahideh Zolfaghari - Dept. of Computer Engineering ImamReza International University Mashhad, Iran
Mehradad Jalali - Dept. of Computer Engineering Islamic Azad University of Mashhad Mashhad, Iran

Ontology matching is a solution for managing heterogeneous information. Various methods have been proposed for matching problem. But matching large ontologiesis a challenge yet. In this paper scalable ontology matching based on partitioning is proposed. The proposed method isconsisted of three stages. At first, input ontologies arepartitioned separately. In second stage anchors are identified in each partition pairs of two ontologies and then by usinganchor distribution, similar partition pairs are identified. At third stage, concepts of each similar partition pairs areclassified based on type of concept. Then scores of importanceof concept are computed and arranged in descending order. Finally, matching is performed on concepts of similarpartitions pair with same concept types. Name similarity metric and WordNet dictionary are used for matchingoperation. Experimental results show that the proposedmethod achieves relatively high precision and it competes with previously proposed partition-based strategies

ontology matching;bottom up hierarchical partitioning;concept type; anchor distribution