Affiliations: State Key Laboratory of Networking and Switching
Technology, Beijing University of Posts and Telecommunications, Beijing,
China | Shenzhen Research Institute, The Chinese University of
Hong Kong, Hong Kong, China | Department of Computer Science and Engineering,
University of Bridgeport, Bridgeport, CT, USA
Note: [] Corresponding author: Shangguang Wang, State Key Laboratory of
Networking and Switching Technology, Beijing University of Posts and
Telecommunications, Box 187, 10 No., Road Xitucheng, Beijing, China. E-mail:
sguang.wang@gmail.com
Abstract: With the popularity of mobile services, an effective context-aware
mobile service adaptation is becoming more and more important for operators. In
this paper, we propose a Co-evolution eXtended Classifier System (CXCS) to
perform context-aware mobile service adaptation. Our key idea is to learn user
context, match adaptation rule, and provide the best suitable mobile services
for users. Different from previous adaptation schemes, our proposed CXCS can
produce a new user's initial classifier population to quicken its converging
speed. Moreover, it can make the current user to predict which service should
be selected, corresponding to an uncovered context. We compare CXCS based on a
common mobile service adaptation scenario with other five adaptation schemes.
The results show the adaptation accuracy of CXCS is higher than 70% on average, and outperforms other schemes.
Keywords: Mobile service, context-aware, service adaptation, learning classifier system, eXtended Classifier System, Co-evolution
