一种基于谓词逻辑的分类规则约简方法  

A Reduction Method of Classification Rules Based On Predicate Logic

作　　者：冀英伟[1] 杨海峰[1] 张继福[1] 

机构地区：[1]太原科技大学计算机科学与技术学院,太原030024

出　　处：《太原科技大学学报》2010年第1期59-64,共6页Journal of Taiyuan University of Science and Technology

基　　金：山西省青年科学基金(2008021028);太原科技大学青年基金(20080315)

摘　　要：利用各种分类规则挖掘方法,提取出的分类规则集中,存在许多的冗余规则,从而降低了分类效率。采用谓词公式描述分类规则,给出了一种分类规则集约简方法。该方法利用谓词公式中的逻辑推理,对分类规则集进行约简,消除了冗余规则。采用恒星光谱数据,实验验证该方法在保证分类规则集的分类能力不变的前提下,有效地提高了分类的效率。There are many redundant rules in classification rule set which is extracted by classification rule mining methods, which results in classification efficiency. A reduction method of classification rules is presented by using the predicate formula to describe classification rules. The method reduces the classification rule set by using logic reasoning in predicate formula, leading to the elimination of the redundant rules. In the end, the experiment results validate that the classification efficiency is improved effectively while classification ability is invariable by celestial spectrum data.

关 键 词：数据挖掘 分类规则 谓词逻辑 冗余规则 恒星光谱数据 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]