Just-in-time recommendation using multi-agents for context-awareness in ubiquitous computing environment

Abstract

Ubiquitous computing is a vision of future in which a computing technology becomes pervasive. One of the most important attributes in ubiquitous computing is context-awareness. The application that the user's contexts are changed continuously over time normally requires a just-in-time recommendation. This paper describes a new just-in-time recommendation method using multi-agents for context-awareness in ubiquitous computing environment. For the just-in-time recommendation, we locally store the recommendation information that the user is likely to need in the near future based on predictions. To predict the information that be expected to be needed by the user, we use data mining methods. We propose a JITR Server Context Agent and a JITR Client Context Agent. They make the just-in-time recommendation continuously with the change of user's contexts and solve the limitation of storage. We explain our method with an example. Several experiments are performed and the experimental results show that our method has a good recommendation performance.