Skip to main content
. Author manuscript; available in PMC: 2012 Dec 1.
Published in final edited form as: Pervasive Mob Comput. 2011 Dec;7(6):746–760. doi: 10.1016/j.pmcj.2011.09.002

Figure 1.

Figure 1

An illustrating example of the clustering algorithm for the recurrent activity discovery; a) three dimensions of accelerometer data, with several occurrences of an activity between t1 and t2. b) The interval coincidence graph, representing the primitives in the time series data. The thickness of edges show higher coincidence between the primitives. c) Primitives with high coincidence are clustered and the recurrent activity is discovered. t1 and t2 are sent to output as the start and end time of the most frequent activity.