This paper has proposed an online classification method
to detect users’ postural actions (e.g., user states, as: sitting,
standing, walking, and running) by using the smartphone
accelerometer sensor. Specifically, a novel DT-based unsupervised
learning method is developed