Towards activity recognition in moving object trajectories from Twitter data

The knowledge about people daily activities is of great value for several application domains. On the one hand, the activity recognition in trajectories has not been deeply investigated. On the other hand, social media data such as tweets can be rich in information about where people go and what they do. We strongly believe that the integration of trajectory data and social media can reveal the activities performed by individuals in daily life. In this paper we propose a new method to infer moving object activities from their trajectories, using knowledge extracted from Twitter data. We evaluate the proposed approach with two datasets and show that it outperforms current works.