CircleSense: A pervasive computing system for recognizing social activities

Social activities have great impact for human beings in psychological health and social relationship. The recognition of social activities can unobtrusively recognize and record users' daily social activities, enabling users to better manage their life. Existing work of social activity recognition focus on recognizing a limited set of social activities and are mainly based on the patterns of individual user such as location pattern, vocal pattern, or others. However, social activities inherently exhibit the patterns with respect to multiple users rather individual user. In this paper, we introduce the concept of social circle, to extract social patterns associated with multiple users in a generic set of social activities. A social circle refers to a set of users frequently gathering to conduct certain social activities. Based on social circle, we design a system called CircleSense that supports accurate recognition of a generic set of social activities. We validate the effectiveness of CircleSense through the real trace collected by 10 volunteers. The result shows that CircleSense outperforms existing methods in terms of accuracy of social activity recognition.