Mining Sensor Streams for Discovering Human Activity Patterns Over Time
Activity discovery and recognition plays an important role in a wide range of applications from assisted living to security and surveillance. Most of the current approaches for activity discovery and recognition rely on a set of predefined activities and assuming a static model of the activities over time. Not only such an assumption violates the dynamic nature of human activities, but also in case of applications such as remote health monitoring and assisted living it will hinder finding useful changes and emerging patterns over time. In this paper, the authors propose a method for finding sequential activity patterns over time from streaming sensor data using multiple time granularity, in the context of smart environment application domain.