摘要
In this paper, the active learning mechanism is proposed to beused in classifier systems to cope with complex problems: an intelligent agent leavesits own signals in the environment and later collects and employs them to assistits learning process. Principles and components of the mechanism are outlined,followed by the introduction of its preliminary implementation in an actual system.An experiment with the system in a dynamic problem is then introduced, togetherwith discussions over its results. The paper is concluded by pointing out somepossible improvements that can be made to the proposed framework.
In this paper, the active learning mechanism is proposed to beused in classifier systems to cope with complex problems: an intelligent agent leavesits own signals in the environment and later collects and employs them to assistits learning process. Principles and components of the mechanism are outlined,followed by the introduction of its preliminary implementation in an actual system.An experiment with the system in a dynamic problem is then introduced, togetherwith discussions over its results. The paper is concluded by pointing out somepossible improvements that can be made to the proposed framework.
