矿业工程专家系统中知识库的质量检测与处理

THE CHECK AND PROCESS OF KNOWLEDGE-BASE QUALITY OF MINING ENGINEERING EXPERT SYSTEM

知识库中知识的质量是矿业工程专家系统性能好坏的决定因素。本文首先分析了矿业工程专家系统知识库可能存在的知识缺陷以及产生这些缺陷的原因,提出了一种基于规则关系树的检测方法和利用知识库运行历史进行检测的方法,可以较完整地检测出知识库中知识可能存在的缺陷,并根据缺陷的类型和产生的部位进行有效处理,这对于开发矿业工程专家系统,提高其性能和知识库的质量都是有益的。

The quality of knowledge in knowledge base is a kernel factor determining whether the mining engineering expert system is of high performances. In this paper, possible knowledge faults of knowledge base in mining engineering expert system and their reasons are analysed at first. Two check methods based on rule-relation tree and running history of knowledge base are proposed. These methods can rather completely check out possible knowledge faults in knowledge base. Some efficient measures are also suggested to process various possible faults occuring in different locations which are very useful to develop expert systems for mining engineering and to improve the performance and quality of their knowledge base.