Design and Implementation of a Context-Aware Guide Application for Mobile Users Based on Machine Learning

This paper proposes a design and implementation of a context-aware application system to guide mobile users about their interesting spots (e.g. restaurants, stores, sightseeing spots) appropriately. A machine learning algorithm enables adaptive recommendation of spots for the mobile users based on their real-time context such as preference, location, weather, time, etc. Our proposed guide system recommends context-aware information for any users by switching two kinds of recommendation algorithms according to the number of user's training data. By experiments using our implemented system in real environments, we confirm that our implemented system correctly works on the off-the-shelf mobile phones having a built-in GPS module and show that it recommends useful information for the mobile users according to their context.