Performance Analysis of Graph-Based Ontology Matching Algorithms
Project no. 756 · Status: Kapanmış · 2016
Summary
Ontologies are domain-specific metadata that describes relationships between properties of a field, sample data, and properties developed for many different purposes. But, they can be different names or properties even within the same domain. However, they can be described in different names or properties even within the same domain. Ontology mapping is a method used to find semantic similarities between existing ontologies and it can be a solution to these differences formed by different names or features. Ontology matching algorithms are used especially for data integration, reuse of ontologies and information extraction. However, rapidly increasing data sizes and the development of information technologies have given rise to the growth of ontologies. Also, ontology matching structures have become more difficult to operate. To overcome this challenge, it has been proposed to divide the ontologies into small pieces and match them into smaller sized structures. In the scope of the proposed project, it is primarily aimed to subdivide large ontologies and then to match these parts with subgraph mining algorithms. Karger Algorithm, which finds on the cut edges of a graph, and clique percolation algorithm, which finds on the cut nodes of a graph, are used to divide the ontologies. Both graph partitioning algorithms, which have not tested on ontologies before, are also yielded successful results. In the next step of the project, these sub-parts obtained by using classical subgraph mining algorithms are matched. GraMi and gSpan algorithms are used as graph mining algorithms. The proposed model is compared with AML, Falcon and another graph-based ontology mapping algorithm in terms of matching accuracy and performance.