원문정보
초록
영어
In order to detect the instant interest in mobile commerce environments, an interactive recommender agent was proposed based on schedules and the algorithm of support vector data description. The agent residing in smart phones obtained the user’s instant interest by comparing the current latitude, longitude and speed of the user’s geographic positions with the preset schedule using support vector data description algorithm. If the current position was in the hypersphere or the tolerant hypersphere, the user were taken as be adherent to the schedule, otherwise, the user would be prompt to confirm whether he or she is adherent or deviated to the schedule. Based on the result of SVDD comparison, a process of determining the behavior states of the user was proposed. And some techniques of interaction between phones and the users were discussed based on vibrations, ringtones, and graphical interfaces. At last, an experiment was conducted to verify the proposed agent, and the result proved that the agent was effective.
목차
1. Introduction
2. The Framework of the Proposed Model
2.1. The process of the framework
2.2. The user’s behavior states
2.3 The algorithm determining the behavior state of the user
3. SVDD
4. Interactions between the User and the Agent
4.1. The user state determination
4.2. Interactions
5. An Experiment and Discussions
5.1 Experimental setup
5.2 Implementation of SVDD algorithm
5.3 Accuracy of detecting behavior deviation
5.4. Resources consumption of the interest agent on mobile phones
5.5 Parameters setup
6. Conclusions and Future Works
Acknowledgements
References