摘要
本文提出的Usex算法,是根据训练集采用解释方法逐步建立决策树,然后生成AI系统知识库所需要的产生式规则集。文中把Usex与其他两个算法作了比较,并进行了许多试验。其结果表明,Usex在时间复杂度、空间复杂度和规则简明性等方面有良好的性能。
Usex,an algorithm,is p roposed in this paper.Based ona training set,Usex builds a decision tree gradually with an explanationprocess,and forms a set of production rules needed for the knowledge baseof the AI system,With several knowledge acquisition algorithms,a varietyof experimentS and their outcomes are introduced here,In conclusion,Usexis better than any other knowledge acquisition algorithms in terms of timecomplexity,space complexity and rule simplicity.
出处
《上海海运学院学报》
1994年第1期38-46,共9页
Journal of Shanghai Maritime University
关键词
决策树
AI系统知识库
Usex算法
explanation,decision trees
rule simplicity,average timefor number of instance sets