The Research Of A Self-Adaptive Agent

Based on machine learning, this paper investigates the intelligent-forming of a self-adaptiveagent-Learner, including how to research, develop and implement the system. Firstly, werestate the story of machine learning and the key factors that influences on the design of learning system. Then, we describe the environment and series of pre-preparations that we need for this learner, and rough set theory, one of the key techniques used in this paper. After that, we improve the generating process of Skowron Discernibility Matrix and give algorithms which simply mine Decision Rules and its Reduction in a decision table. By using these algorithms, we design this intelligently self-adaptive agent, and implement it based on VC combined with Oracle 9i. Finally, we list some testing data which validate the correctness and feasibility of this system.