摘要
数据库通常包含很多冗余特征,找出重要特征叫做特征提取。本文提出一种基于属性重要度的启发式特征选取算法。该算法以属性重要度为迭代准则得到属性集合的最小约简。
A database always contains many attributes （sometimes instead of feature ）that are redundant and not necessary for rule discovery . Feature selection is to find optimal feature subset. In this paper, we introduce an algorithm which is using the importance of attribute with heuristics for feature selection. This algorithm takes the importance of attribute as the iterative criterion and 5nds the least reduction of attributeset.
出处
《自动化与仪器仪表》
2005年第5期13-14,17,共3页
Automation & Instrumentation
基金
四川省科技厅应用基础项目基金资助(03JY029-018-1)
关键词
粗集
重要度
特征选取
Rough set
Importance
Feature selection
