西南交通大学学报
西南交通大學學報
서남교통대학학보
JOURNAL OF SOUTHWEST JIAOTONG UNIVERSITY
2010年
2期
312-316
,共5页
云变换%粗糙集%决策树%加权平均粗糙度%特性关系
雲變換%粗糙集%決策樹%加權平均粗糙度%特性關繫
운변환%조조집%결책수%가권평균조조도%특성관계
cloud transform%rough set%decision tree%weighted mean roughness%characteristicrelatiOn
为了从不完备信息系统中挖掘分类知识,提出了一种基于云变换和特性关系下粗糙集的决策树构造算法(DTCCRSCR).其核心思想是,利用云变换离散化连续型数据,选择特性关系下加权平均粗糙度最小的属性作为当前的分裂节点.实验表明,由DTCCRSCR构造的决策树不仅结构简单,分类准确率较高,而且分类规则也容易理解.
為瞭從不完備信息繫統中挖掘分類知識,提齣瞭一種基于雲變換和特性關繫下粗糙集的決策樹構造算法(DTCCRSCR).其覈心思想是,利用雲變換離散化連續型數據,選擇特性關繫下加權平均粗糙度最小的屬性作為噹前的分裂節點.實驗錶明,由DTCCRSCR構造的決策樹不僅結構簡單,分類準確率較高,而且分類規則也容易理解.
위료종불완비신식계통중알굴분류지식,제출료일충기우운변환화특성관계하조조집적결책수구조산법(DTCCRSCR).기핵심사상시,이용운변환리산화련속형수거,선택특성관계하가권평균조조도최소적속성작위당전적분렬절점.실험표명,유DTCCRSCR구조적결책수불부결구간단,분류준학솔교고,이차분류규칙야용역리해.
A new algorithm,DTCCRSCR(decision tree construction based on cloud transform and rough set theory under characteristic relation),Was proposed for mining classification knowledge from incomplete information systems.It utilizes cloud transform to discretize continuous data,and then the attribute with the smallest weighted mean roughness under the characteristic relation is selected as the current splitting node.Experimental results show that decision trees constructed by the DTCCRSCR have a simpler structure,a higher classification accuracy and more undemmndable rules than those based on the C5.0 algorithm.
