Context and Activity Recognition for Personalized Mobile Recommendations.

Through the use of mobile devices, contextual information about users can be derived to use as an additional information source for traditional recommendation algorithms. This paper presents a framework for detecting the context and activity of users by analyzing sensor data of a mobile device. The recognized activity and context serves as input for a recommender system, which is built on top of the framework. Through context-aware recommendations, users receive a personalized content offer, consisting of relevant information such as points-of-interest, train schedules, and touristic info. An evaluation of the recommender system and the underlying context-recognition framework demonstrates the impact of the response times of external information providers. The data traffic on the mobile device required for the recommendations shows to be limited. A user evaluation confirms the usability and attractiveness of the recommender. The recommendations are experienced as effective and useful for discovering new venues and relevant information.