摘要
针对属性过多对于有效的数据挖掘很不利以及约简中差别矩阵的产生会占用较大存储空间的问题,提出了一种基于粗糙集和信息增益的属性约简改进算法。该算法首先采用信息增益技术对决策表属性进行相关分析,删除部分冗余属性,减小属性约简的复杂度,然后直接从决策表中提取出分明函数,求出属性约简。由于避免了分明矩阵的生成,因此该算法不仅节约了时间和空间,而且提高了效率。
Aiming at the problems of too many attributes in data mining and much space acquired while generating the discernibility ma-trix,an improved algorithm for attribute reduction which is based on the rough sets and information gain,is put forward.The analysis of information gain technology is used to analyze the relationship between attributes to reduce the complexity of reduction.We can get the attribute reduction without generating the discernibility matrix.Less time and space complexity are acquired.And it is verified that the algorithm is effective.
出处
《计算机工程与设计》
CSCD
北大核心
2009年第24期5698-5700,共3页
Computer Engineering and Design
基金
国家自然科学基金项目(60873196)
甘肃省科技攻关基金项目(2GS064-AS2-035-03)
关键词
粗糙集
信息增益
属性约简
值约简
分明矩阵
rough sets
information gain
attributes reduction
value reduction
discernibility matrix
