Reduction techniques for instance-based learning algorithms

被引:904
作者
Wilson, DR [1 ]
Martinez, TR [1 ]
机构
[1] Brigham Young Univ, Dept Comp Sci, Neural Network & Machine Learning Lab, Provo, UT 84602 USA
关键词
instance-based learning; nearest neighbor; instance reduction; pruning; classification;
D O I
10.1023/A:1007626913721
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Instance-based learning algorithms are often faced with the problem of deciding which instances to store for use during generalization. Storing too many instances can result in large memory requirements and slow execution speed, and can cause an oversensitivity to noise. This paper has two main purposes. First, it provides a survey of existing algorithms used to reduce storage requirements in instance-based learning algorithms and other exemplar-based algorithms. Second, it proposes six additional reduction algorithms called DROP1-DROP5 and DEL (three of which were first described in Wilson & Martinez, 1997c, as RT1-RT3) that can be used to remove instances from the concept description. These algorithms and 10 algorithms from the survey are compared on 31 classification tasks. Of those algorithms that provide substantial storage reduction, the DROP algorithms have the highest average generalization accuracy in these experiments, especially in the presence of uniform class noise.
引用
收藏
页码:257 / 286
页数:30
相关论文
共 54 条
[1]   INSTANCE-BASED LEARNING ALGORITHMS [J].
AHA, DW ;
KIBLER, D ;
ALBERT, MK .
MACHINE LEARNING, 1991, 6 (01) :37-66
[2]   TOLERATING NOISY, IRRELEVANT AND NOVEL ATTRIBUTES IN INSTANCE-BASED LEARNING ALGORITHMS [J].
AHA, DW .
INTERNATIONAL JOURNAL OF MAN-MACHINE STUDIES, 1992, 36 (02) :267-287
[3]  
[Anonymous], MACHINE LEARNING /
[4]  
[Anonymous], MACH LEARN INT WORKS
[5]  
Batchelor B.G., 1978, PATTERN RECOGNITION
[6]  
BIBERMAN Y, 1994, P 9 EUR C MACH LEARN, P49
[7]  
Brodley C. E., 1993, P 10 INT C MACH LEAR, P17, DOI DOI 10.1016/B978-1-55860-307-3.50009-5
[8]  
Broomhead D. S., 1988, Complex Systems, V2, P321
[9]  
CAMERONJONES RM, 1995, P 8 AUSTR JOINT C AR, P99
[10]   A MASSIVELY PARALLEL ARCHITECTURE FOR A SELF-ORGANIZING NEURAL PATTERN-RECOGNITION MACHINE [J].
CARPENTER, GA ;
GROSSBERG, S .
COMPUTER VISION GRAPHICS AND IMAGE PROCESSING, 1987, 37 (01) :54-115