Service Composition Techniques Using Data Mining for Ubiquitous Computing Environments

When users want to have a suitable application in ubiquitous computing environments, it is necessary to discover and compose primitive services considering context information. However, previous works focused on only service discovery for user queries and lack the consideration of context information such as location, time, network, computing environment, and preference for users. They also do not use the service history information on service composition. Therefore, we present a framework for a service provisioning middleware system that can discover primitive services and compose dynamic complex services according to the context information. We also describe an algorithm of service composition which uses the service history information and an ontology engine with data mining. Finally, we show that our experiments enhance the possibility of provisioning services considering user's preference and thus provide users with newly composed optimal services.