Estimation of interruptibility during office work based on PC activity and conversation

Abstract

The chances of being interrupted by online communication systems, such as email, instant messenger, and micro-blog, are rapidly increasing. For the adequate control of interruption timing, the real-time estimation of the interruptibility of the user is required. In this study, we propose an interruptibility estimation method using PC activity and conversational voice detection based on the wavelet transform. The offline estimation was applied to a dataset of 50 hours obtained from 10 users. The results indicated the feasibility of improving the interruptibility estimation accuracy by the automatic detection of the existence and end of conversations. © 2013 Springer-Verlag Berlin Heidelberg.