Efficient Computer-Mediated Communication via Automatic Context Acquisition
DescriptionThe aim of this thesis is to develop a system that supports the effectiveness communication of remote users by avoiding disruptive interruptions on contactee’s side and by supporting quick and useful responses on the contacter’s side. Prior work concentrated on the contactee to avert interruptions being disruptive, but interruptions also can be advantageous for both communication partners. So for an effective and efficient interruption handling the needs of both participants have to be considered. That is the main focus of this thesis.
The needs and preferences of the users are determined through the automated analysis of their context by sensors. On the contactee's side the focus is on her availability. Investigated factors of the contacter's context are the subject or content of her message along with the urgency and importance. The relationship is a further aspect, which is taken additionally into account. Out of the sensor data, the system estimates the general availability of the contactee for incoming messages and the subject, importance and urgency of the contacter’s message. On the basis of the known preferences of both users the system estimates the effectiveness of an upcoming interruption and decides whether the interruption will occur or not and in which way it will take place. For the matching of the preferences and the calculation of the effectiveness a Bayesian network is used. Further, the results and the system itself will be tested and validated to verify the appropriateness of this approach and will be compared with the user-driven approach for effective interruption handling.
PeopleTom Gross, Supervisor
Simone Braun, Master's candidate