化石鉴定专家系统中具学习功能决策树的研究与实现

Study on Learning Decision Tree in Expert System and Its Implementation

决策树应用于专家系统的知识表示可以实现决策过程的可视化。本文详细地给出了决策树的两种学习算法,实现了决策树中知识的增长以及从误判中学习新的知识的功能。此方法在我们开发的一个可视化专家系统工具(Visual ESTool)中得到了应用。最后给出了用该方法演示了建立一个决策树的详细过程。

Decision tree can visualize the decision process in representing the knowledge of an expert system. This paper presents two learning algorithms which can implement the ability of increasing knowledge in decision tree and the ability of learning from misjudgement.Based on the algorithms,a kind of knowledge base used in Visual ESTool is established and an example demonstrating the process of establishing a decision tree is presented in this paper.