header

Humans, Semantic Services and Similarity: A User Study of Semantic Web Services Matching and Composition

32 Pages Posted: 8 Jul 2018 Publication Status: Accepted

See all articles by Eran Toch

Eran Toch

Carnegie Mellon University

Iris Reinhartz-Berger

University of Haifa

Dov Dori

Technion-Israel Institute of Technology

Abstract

Inferring similarity between Web services is a fundamental construct for service matching and composition. However, there is little evidence of how humans perceive similarity between services, a crucial knowledge for designing usable and practical service matching and composition algorithms. In this study we have experimented with 127 users to define and evaluate a model for service similarity in the context of semantic Web services. Our findings show that humans take a complex and sophisticated approach towards service similarity, which is more fine-grained than suggested by theoretical models of service similarity, such as logic-based approaches. We define a similarity model, based on our empirical findings and prove that the similarity model, expressed by a distance metric, is complete and that it closely predicts humans' perceptions of service similarity. Finally, we describe an application of a Web service search engine that implements our model.

Keywords: Semantic Web services, User study, Similarity, Approximation

Suggested Citation

Toch, Eran and Reinhartz-Berger, Iris and Dori, Dov, Humans, Semantic Services and Similarity: A User Study of Semantic Web Services Matching and Composition (2011). Available at SSRN: https://ssrn.com/abstract=3199503 or http://dx.doi.org/10.2139/ssrn.3199503

Eran Toch (Contact Author)

Carnegie Mellon University ( email )

Pittsburgh, PA 15213-3890
United States

Iris Reinhartz-Berger

University of Haifa ( email )

Mount Carmel
Haifa, 31905
Israel

Dov Dori

Technion-Israel Institute of Technology ( email )

Technion City
Haifa 32000, Haifa 32000
Israel

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
348
Downloads
8
PlumX Metrics