In this paper, a novel layered clustering approach is proposed to cluster web services in order to facilitate web service selection process. The process of web service selection from a rapidly growing number of functionally similar services in the internet, results in an increase of service discovery cost, transforming data time between services and service searching time. Though suitable technologies for web services clustering are being developed, blending neural networks and swarm-based algorithms is not prevailing. A novel two-phase clustering approach involving ART (Adaptive Resonance Theory) network for primary clustering with functional data and swarm algorithms (BOIDs, ABC and PSO) for sub-clustering with non-functional data (metadata, QoS and service-generated data respectively) is proposed in this work. As a result of this layered approach, the computational overhead is greatly reduced and the search space is also abridged significantly in order to obtain optimal services.
Digital Object Identifier (DOI)
Praveen Joe, I.R. and Varalakshmi, P.
"An Analysis on Web-Service-Generated Data to Facilitate Service Retrieval,"
Applied Mathematics & Information Sciences: Vol. 13:
1, Article 7.
Available at: https://digitalcommons.aaru.edu.jo/amis/vol13/iss1/7