Social networks : service selection and recommendation
JA Al-Sharawneh · UTS ePRESS (University of Technology Sydney) · 2012
The Service-Oriented Computing paradigm is widely acknowledged for its potential to revolutionize the world of computing through the utilization of Web services.It is expected that Web services will fully leverage the Semantic Web to outsource some of their functionalities to other Web services that provide value-added services, and by integrating the business logic of Web services in the form of business to business and business to consumer e-commerce applications.In the Service Web, Web services and Web-Based Social Networks are emerging in which a wide range of similar functionalities are expected to be offered by a vast number of Web services, and applications can search and compose services according to users' needs in a seamless and an automatic fashion.Web services are expected to outsource some of their functionalities to other Web services.In such situations, some services may be new to the service market, and some may act maliciously in order to be selected.A key requirement is to provide mechanisms for quality selection and recommendation of relevant Web services with perceived risk considerations.Although the future of Web service selection and recommendation looks promising, there are challenging issues related to user knowledge and behavior, as well as issues related to recommendation approaches.This dissertation addresses the demanding issues in Web service selection and recommendation from theory and practice perspectives.These challenges include cold-start users, who represent more than 50% of the social network population, the capture of users' preferences, risk mitigation in service selection, customers' privacy and application scalability.This dissertation proposes a novel approach to automate social-based Web service selection and recommendation in a dynamic environment.It utilizes Web-Based Social Networks and the "Follow the Leader" strategy, for a Credibility-based framework that includes two credibility models: the user Credibility model which is used to qualify without the support of the following people: First and foremost, I would like to express my gratitude to my supervisor Professor Mary-Anne Williams for her wise advice, careful guidance, useful connections, constant support and enduring patience throughout my studies and while writing this dissertation.Mary-Anne's advice on the use of "Follow the Leader" strategy was the turning point of my research.Learning how to research and produce publications is much easier with a supervisor who has as much experience and enthusiasm as Mary-Anne.I would like also to