RIMER中学生学习专家系统的开发与设计

Development and Design of RIMER Learning Expert System

RIMER中学生学习专家系统是采用证据推理方法的新型学习型专家系统,实现了中学生有效的自我学习与评估,促进了中学生学习效率与质量的提高。在介绍RIMER中学生学习专家系统的总体结构基础上,阐述了RIM-ER中学生学习专家系统中的输入输出、正向推理、学习训练等各个模块的详细设计,并用共轭梯度算法实现了对置信库规则的训练学习。实验结果证明该算法高效实用,克服了Yang等给出的置信规则库在速度、简易性、收敛性等方面的诸多局限性。

RIMER middle school students learning expert system is a kind of evidence reasoning method new learning ex- pert system, which realizes effective self-learning and assessment for the middle school students, and promote the improve- ment of the efficiency and quality of middle school students. This paper introduces the expert system on the basis of the general architecture, described in the RIMER expert system of input and output, forward reasoning, learning the various modules of the detailed design, and the conjugate gradient algorithm is used to realize the learning. Experiment result to the confidence library rules proves that the algorithm is efficient and practical, which overcomes the rule base in speed, convergence, simplicity and limitations etc.