The Community for Technology Leaders
Green Image
Issue No. 03 - May-June (2015 vol. 8)
ISSN: 1939-1374
pp: 439-452
Jinhui Yao , Palo Alto Research Center (PARC), Webster
Wei Tan , IBM T.J. Watson Research Center, New York
Surya Nepal , CSIRO Computational Informatics (CCI), Sydney, NSW, Australia
Shiping Chen , CSIRO Computational Informatics (CCI), Sydney, NSW, Australia
Jia Zhang , Carnegie Mellon University-Silicon Valley, San Francisco
David De Roure , Oxford e-Research Centre, Oxford University, Oxford, United Kingdom
Carole Goble , School of Computer Science, University of Manchester, United Kingdom
In the paradigm of service oriented science, scientific computing applications and data are all wrapped as web accessible services. Scientific workflows further integrate these services to answer complex research questions. However, our earlier study conducted on myExperiment has revealed that although the sharing of service-based capabilities opens a gateway to resource reuse, in practice, the degree of reuse is very low. This finding has motivated us to propose ServiceMap to provide navigation facility through the network of services to facilitate the design and development of scientific workflows. This paper proposes ReputationNet as an enhancement of ServiceMap, to incorporate the often-ignored reputation aspects of services/workflows and their publishers, in order to offer better service and workflow recommendations. We have developed a novel model to reflect the reputation of e-Science services/workflows, and developed heuristic algorithms to provide service recommendations based on reputations. Experiments on myExperiment have illustrated a strong positive correlation (with Pearson correlation coefficient 0.82) between the reputation scores computed and the actual performance (i.e. usage frequency) of the services/workflows, which demonstrates the effectiveness of our approach.
Educational institutions, Electronic mail, Communities, Correlation, Computational modeling, Social network services, Service-oriented architecture

J. Yao et al., "ReputationNet: Reputation-Based Service Recommendation for e-Science," in IEEE Transactions on Services Computing, vol. 8, no. 3, pp. 439-452, 2015.
260 ms
(Ver 3.1 (10032016))