Inferring user activities from spatial-temporal data in mobile phones

Abstract

Activity inference is a key to the development of various ubiquitous computing applications. Here, we observe that users perform several actions in their mobile phone: take photos, perform check-in, and access Wi-Fi networks. These behaviors generate spatial-temporal data that could be utilized to capture user activities. Hence, three features are extracted for activities inference: 1) geographical feature: indicating where user performs activities; 2) temporal feature: indicating when user performs activities; and 3) semantic feature: showing semantic concept of a place from location-based social networks. Here, we propose Spatial-Temporal Activity Inference Model (STAIM) to infer users` activities from aforementioned features. Experimental results show that STAIM is able to effectively infer users` activities, achieving 75% accuracy on average. Copyright 2015 © ACM.