An architecture for user preference-based IoT service selection in cloud computing using mobile devices for smart campus

Master Thesis

2015

Permanent link to this Item
Authors
Journal Title
Link to Journal
Journal ISSN
Volume Title
Publisher
Publisher

University of Cape Town

License
Series
Abstract
The Internet of things refers to the set of objects that have identities and virtual personalities operating in smart spaces using intelligent interfaces to connect and communicate within social environments and user context. Interconnected devices communicating to each other or to other machines on the network have increased the number of services. The concepts of discovery, brokerage, selection and reliability are important in dynamic environments. These concepts have emerged as an important field distinguished from conventional distributed computing by its focus on large-scale resource sharing, delivery and innovative applications. The usage of Internet of Things technology across different service provisioning environments has increased the challenges associated with service selection and discovery. Although a set of terms can be used to express requirements for the desired service, a more detailed and specific user interface would make it easy for the users to express their requirements using high-level constructs. In order to address the challenge of service selection and discovery, we developed an architecture that enables a representation of user preferences and manipulates relevant descriptions of available services. To ensure that the key components of the architecture work, algorithms (content-based and collaborative filtering) derived from the architecture were proposed. The architecture was tested by selecting services using content-based as well as collaborative algorithms. The performances of the algorithms were evaluated using response time. Their effectiveness was evaluated using recall and precision. The results showed that the content-based recommender system is more effective than the collaborative filtering recommender system. Furthermore, the results showed that the content-based technique is more time-efficient than the collaborative filtering technique.
Description

Reference:

Collections