Test strategies for cost-sensitive decision trees

被引:89
作者
Ling, Charles X. [1 ]
Sheng, Victor S.
Yang, Qiang
机构
[1] Univ Western Ontario, Dept Comp Sci, London, ON N6A 5B7, Canada
[2] Hong Kong Univ Sci & Technol, Dept Comp Sci, Hong Kong, Hong Kong, Peoples R China
基金
加拿大自然科学与工程研究理事会;
关键词
induction; concept learning; mining methods and algorithms; classification;
D O I
10.1109/TKDE.2006.131
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
In medical diagnosis, doctors must often determine what medical tests ( e. g., X-ray and blood tests) should be ordered for a patient to minimize the total cost of medical tests and misdiagnosis. In this paper, we design cost-sensitive machine learning algorithms to model this learning and diagnosis process. Medical tests are like attributes in machine learning whose values may be obtained at a cost ( attribute cost), and misdiagnoses are like misclassifications which may also incur a cost ( misclassification cost). We first propose a lazy decision tree learning algorithm that minimizes the sum of attribute costs and misclassification costs. Then, we design several novel "test strategies" that can request to obtain values of unknown attributes at a cost ( similar to doctors' ordering of medical tests at a cost) in order to minimize the total cost for test examples ( new patients). These test strategies correspond to different situations in real-world diagnoses. We empirically evaluate these test strategies, and show that they are effective and outperform previous methods. Our results can be readily applied to real-world diagnosis tasks. A case study on heart disease is given throughout the paper.
引用
收藏
页码:1055 / 1067
页数:13
相关论文
共 20 条
[11]  
LIZOTTE D, 2003, P 19 C UNC ART INT
[12]  
MELVILLE P, 2004, P WORKSH UT BAS DAT
[13]  
MELVILLE P, 2004, P 4 INT C DAT MIN
[14]  
NUNEZ M, 1991, MACH LEARN, V6, P231, DOI 10.1007/BF00114778
[15]  
Quinlan J. R., 1993, C4 5 PROGRAMS MACHIN
[16]  
SHENG VS, 2006, P 21 NAT C ART INT
[17]   COST-SENSITIVE LEARNING OF CLASSIFICATION KNOWLEDGE AND ITS APPLICATIONS IN ROBOTICS [J].
TAN, M .
MACHINE LEARNING, 1993, 13 (01) :7-33
[18]  
TING KM, 1998, P 2 EUR S PRINC DAT, P23
[19]   Cost-Sensitive Classification: Empirical Evaluation of a Hybrid Genetic Decision Tree Induction Algorithm [J].
Turney, Peter D. .
JOURNAL OF ARTIFICIAL INTELLIGENCE RESEARCH, 1994, 2 :369-409
[20]  
ZUBEK VB, 2002, P 19 INT C MACH LEAR, P27