A Hybrid Semantic Similarity Measuring Approach For Annotating Wsdl Documents With Ontology Concepts

Keywords

Ontology; Semantic annotation; Semantic similarity; Semantic Web Services; Web Services

Abstract

Semantic annotation of legacy Web Services is one of the fast and efficient ways to implement Semantic Web Service paradigm. Semantic similarity between concepts in WSDL (Web Services Description Language) document and ontology concepts is the backbone of semantic annotation of legacy Web Services. The overwhelming majority of previous works focused mainly on semantic similarity of concepts in a specific domain ontology. However, the concepts used in Web Services are often from multiple sources or different domain ontologies. This makes traditional approaches no longer applicable. To address this, we propose a hybrid measuring approach to measure semantic similarity between concepts in WSDL documents and concepts in OWL (Ontology Web Language) files. The proposed approach mainly consists of two parts: lexical-level similarity measuring and structural-level similarity measuring. Specially, we fusion adopt three commonly used approaches, i.e., edge-based, feature-based, and information content-based semantic similarity measuring approaches. Specifically, we map the above mentioned three approaches to three proposed internal features, i.e., depth, width, and density, in the abstract tree structure when measuring structural-level similarity. We conduct experimental comparisons, and the results show that the proposed approach provides better accuracy. Furthermore, the proposed approach can be applied in any user defined Web Services description documents in theory with a wider range of application.

Publication Date

1-1-2017

Publication Title

International Journal of Innovative Computing, Information and Control

Volume

13

Issue

4

Number of Pages

1221-1242

Document Type

Article

Personal Identifier

scopus

Socpus ID

85025088769 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/85025088769

This document is currently not available here.

Share

COinS