Multi-Algorithm Ontology Mapping With Automatic Weight Assessment And Background Knowledge
Loading...
Date
2015-06
Authors
NARANJAN SINGH, SHAILENDRA SINGH
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Ontologies, being the backbone of the Semantic Web, have been used to facilitate standardisation of terminologies across different applications. However, the development of various ontologies has in itself resulted in the problem of heterogeneity due to different terminologies used to build different ontologies. To address the problem of ontology heterogeneity, ontology mapping has been viewed as a solution to help discover similar elements from different ontologies. Single ontology mapping algorithms have deficiencies of not being able to cover different properties of ontologies. Although multi-algorithm approaches performed better, the key issue is in the selection of algorithms for a particular ontology mapping case. This thesis presents a two-phase multi-algorithm ontology mapping approach. Phase 1 provides a mechanism to suggest the right combination of single ontology mapping algorithms for every unique problem separately. Phase 2 aggregates the combination of single ontology mapping algorithms followed by a background knowledge look-up for mappings which were missed by the single ontology mapping algorithms. Experiments using the OAEI (Ontology Alignment Evaluation Initiative) dataset resulted in better precision, recall and F-measure in comparison to existing ontology mapping algorithms. In the case of mapping the Anatomy ontology, the proposed approach ranked first for recall and F-measure, and fourth in precision against 20 other algorithms submitted in OAEI for the year 2013. The proposed approach has opened up possibilities for future work on large-scale matching evaluation.
Description
Keywords
Multi-Algorithm Ontology Mapping With Automatic Weight Assessment , And Background Knowledge