摘要： 归纳学习是解决知识自动获取的有效方法。针对ID3算法、基于粗集的归纳学习以及其它一些归纳学习方法存在的问题,提出了一种新的归纳学习算法ITIL。此算法用信息增益为启发式,选择尽量少的重要属性或组合,以可分辨性为依据提取规则。许多实例表明,这些规则不仅简单,而且冗余小。作为知识获取模块的一部分,ITIL已被集成到一个“基于知识发现的医疗诊断辅助系统”动态知识库子系统中。

关键词: 归纳学习, 粗集, 信息论

Abstract: Inductive iearning is a useful method for solving the automatic knowledge acquisition in intelligent systems . In relation to problems in some inductive learning algorithms such as ID3, learning based on rough set and RITIO proposed in Ref. 6, a new inductive learning method ITIL(Information Theory-based Induction Learning) is put forward , in which fewer important attributes discerning objects in decision tables are choosen to extract rules based on information entropy and rough set theory. Some examples show that acquired rules using the proposed algorithm are both simple and including less redundancy compared with a before-mentioned algorithm . As a part of knowledge acquisition module, ITIL has been integrated into a medical diagnosis aided system .

中图分类号: 

• TP182