Optimal reference subset selection for nearest neighbor classification by tabu search

被引:41
作者
Zhang, HB [1 ]
Sun, GY [1 ]
机构
[1] Beijing Polytech Univ, Comp Inst, Beijing 100044, Peoples R China
基金
中国国家自然科学基金;
关键词
nearest neighbor classifications; tabu search; reference set; prototype selection;
D O I
10.1016/S0031-3203(01)00137-6
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
This paper presents an approach to select the optimal reference subset (ORS) for nearest neighbor classifier. The optimal reference subset, which has minimum sample size and satisfies a certain resubstitution error rate threshold, is obtained through a tabu search (TS) algorithm. When the error rate threshold is set to zero, the algorithm obtains a near minimal consistent subset of a given training set, While the threshold is set to a small appropriate value. the obtained reference Subset may have reasonably good generalization capacity. A neighborhood exploration method and an aspiration criterion are proposed to improve the efficiency of TS. Experimental results based on a number of typical data sets are presented and analyzed to illustrate the benefits of the proposed method. The performances of the result consistent and non-consistent reference subsets are evaluated. (C) 2002 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.
引用
收藏
页码:1481 / 1490
页数:10
相关论文
共 17 条
[1]  
[Anonymous], 1997, TABU SEARCH
[2]  
Cerverón V, 1998, LECT NOTES COMPUT SC, V1518, P248
[3]   FINDING PROTOTYPES FOR NEAREST NEIGHBOR CLASSIFIERS [J].
CHANG, CL .
IEEE TRANSACTIONS ON COMPUTERS, 1974, C 23 (11) :1179-1184
[4]  
Dasarathy B. V., 1991, IEEE COMPUT SOC TUTO
[5]   MINIMAL CONSISTENT SET (MCS) IDENTIFICATION FOR OPTIMAL NEAREST-NEIGHBOR DECISION SYSTEMS-DESIGN [J].
DASARATHY, BV .
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS, 1994, 24 (03) :511-517
[6]  
Devijver P. A., 1980, Proceedings of the 5th International Conference on Pattern Recognition, P72
[7]   BAYES ERROR ESTIMATION USING PARZEN AND K-NN PROCEDURES [J].
FUKUNAGA, K ;
HUMMELS, DM .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1987, 9 (05) :634-643
[8]   BIAS OF NEAREST NEIGHBOR ERROR-ESTIMATES [J].
FUKUNAGA, K ;
HUMMELS, DM .
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, 1987, 9 (01) :103-112
[9]  
Fukunaga K., 1990, INTRO STAT PATTERN R
[10]  
GATES GW, 1972, IEEE T INFORM THEORY, V18, P431, DOI 10.1109/TIT.1972.1054809