Rank Entropy-Based Decision Trees for Monotonic Classification

被引:148
作者
Hu, Qinghua [1 ]
Che, Xunjian [1 ]
Zhang, Lei [2 ]
Zhang, David [2 ,3 ]
Guo, Maozu [1 ]
Yu, Daren [1 ]
机构
[1] Harbin Inst Technol, Harbin 150001, Peoples R China
[2] Hong Kong Polytech Univ, Dept Comp, Hong Kong, Hong Kong, Peoples R China
[3] Hong Kong Polytech Univ, Biometr Technol Ctr UGC CRC, Hong Kong, Hong Kong, Peoples R China
基金
中国国家自然科学基金;
关键词
Monotonic classification; rank entropy; rank mutual information; decision tree; MULTIATTRIBUTE UTILITY-THEORY; ROUGH APPROXIMATION; METHODOLOGY; SETS;
D O I
10.1109/TKDE.2011.149
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In many decision making tasks, values of features and decision are ordinal. Moreover, there is a monotonic constraint that the objects with better feature values should not be assigned to a worse decision class. Such problems are called ordinal classification with monotonicity constraint. Some learning algorithms have been developed to handle this kind of tasks in recent years. However, experiments show that these algorithms are sensitive to noisy samples and do not work well in real-world applications. In this work, we introduce a new measure of feature quality, called rank mutual information (RMI), which combines the advantage of robustness of Shannon's entropy with the ability of dominance rough sets in extracting ordinal structures from monotonic data sets. Then, we design a decision tree algorithm (REMT) based on rank mutual information. The theoretic and experimental analysis shows that the proposed algorithm can get monotonically consistent decision trees, if training samples are monotonically consistent. Its performance is still good when data are contaminated with noise.
引用
收藏
页码:2052 / 2064
页数:13
相关论文
共 41 条
[11]   Growing decision trees in an ordinal setting [J].
Cao-Van, K ;
De Baets, B .
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS, 2003, 18 (07) :733-750
[12]   Fuzzy SLIQ decision tree algorithm [J].
Chandra, B. ;
Varghese, P. Paul .
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS PART B-CYBERNETICS, 2008, 38 (05) :1294-1301
[13]  
da Costa JP, 2005, LECT NOTES ARTIF INT, V3720, P690, DOI 10.1007/11564096_70
[14]   MULTIPLE CRITERIA DECISION-MAKING, MULTIATTRIBUTE UTILITY-THEORY - THE NEXT 10 YEARS [J].
DYER, JS ;
FISHBURN, PC ;
STEUER, RE ;
WALLENIUS, J ;
ZIONTS, S .
MANAGEMENT SCIENCE, 1992, 38 (05) :645-654
[15]  
Feelders A, 2003, LECT NOTES COMPUT SC, V2810, P1, DOI 10.1007/978-3-540-45231-7_1
[16]  
Feelders Ad, 2010, Proceedings 2010 10th IEEE International Conference on Data Mining (ICDM 2010), P803, DOI 10.1109/ICDM.2010.92
[17]  
Frank E., 2001, EUR C MACH LEARN, P145, DOI [10.1007/3-540-44795-413, DOI 10.1007/3-540-44795-413]
[18]   Rough approximation by dominance relations [J].
Greco, S ;
Matarazzo, B ;
Slowinski, R .
INTERNATIONAL JOURNAL OF INTELLIGENT SYSTEMS, 2002, 17 (02) :153-171
[19]   Rough approximation of a preference relation by dominance relations [J].
Greco, S ;
Matarazzo, B ;
Slowinski, R .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 1999, 117 (01) :63-83
[20]   Rough sets methodology for sorting problems in presence of multiple attributes and criteria [J].
Greco, S ;
Matarazzo, B ;
Slowinski, R .
EUROPEAN JOURNAL OF OPERATIONAL RESEARCH, 2002, 138 (02) :247-259