Algorithms for subsetting attribute values with Relief

被引：11
作者：
Demsar, Janez [1 ]
机构：
[1] Univ Ljubljana, Fac Comp & Informat Sci, Ljubljana, Slovenia
来源：
MACHINE LEARNING | 2010年 / 78卷 / 03期
关键词：
Machine learning; Attribute quality estimation; Relief;
D O I：
10.1007/s10994-009-5164-0
中图分类号：
TP18 [人工智能理论];
学科分类号：
081104 ; 0812 ; 0835 ; 1405 ;
摘要：
Relief is a measure of attribute quality which is often used for feature subset selection. Its use in induction of classification trees and rules, discretization, and other methods has however been hindered by its inability to suggest subsets of values of discrete attributes and thresholds for splitting continuous attributes into intervals. We present efficient algorithms for both tasks.
引用
收藏
页码：421 / 428
页数：8
相关论文
共 7 条
[1]   EFFICIENT ALGORITHMS FOR AGGLOMERATIVE HIERARCHICAL-CLUSTERING METHODS [J].
DAY, WHE ;
EDELSBRUNNER, H .
JOURNAL OF CLASSIFICATION, 1984, 1 (01) :7-24
[2]  
Jakulin A, 2005, THESIS U LJUBLJANA
[3]  
KIRA K, 1992, AAAI-92 PROCEEDINGS : TENTH NATIONAL CONFERENCE ON ARTIFICIAL INTELLIGENCE, P129
[4]  
Kononenko I., 1994, LECT NOTES COMPUT SC, V784, P171, DOI [10.1007/3-540-57868-4_57/COVER, DOI 10.1007/3-540-57868-457, 10.1007/3-540-57868-4_57, DOI 10.1007/3-540-57868-4_57]
[5]  
Kononenko I, 2008, CH CRC DATA MIN KNOW, P169
[6]  
KRAMER S, 1994, P ML COLT 94 WORKSH
[7]   Theoretical and empirical analysis of ReliefF and RReliefF [J].
Robnik-Sikonja, M ;
Kononenko, I .
MACHINE LEARNING, 2003, 53 (1-2) :23-69
1