ENERGY FUNCTIONS FOR MINIMIZING MISCLASSIFICATION ERROR WITH MINIMUM-COMPLEXITY NETWORKS

被引:23
作者
TELFER, BA
SZU, HH
机构
关键词
CLASSIFICATION; CONJUGATE GRADIENT; ENERGY FUNCTIONS; GRADIENT DESCENT; MINIMUM MISCLASSIFICATION ERROR; NEURAL NETWORKS; OBJECTIVE FUNCTIONS;
D O I
10.1016/0893-6080(94)90102-3
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
For automatic target recognition, a neural network is desired that minimizes the number of misclassifications with the minimum network complexity. Minimizing network complexity is important for both improving generalization and simplifying implementation. The least mean squares (LMS) energy function used in standard back propagation does not always produce such a network. Therefore, two minimum misclassification error (MME) energy functions are advanced to achieve this. Examples are given in which LMS requires five times as many hidden units in a multilayer perceptron to achieve test set classification accuracy similar to that achieved with the MME functions. Examples are given to provide insight into the nature of the LMS performance, namely that LMS approximates the a posteriori probabilities and class boundaries emerge indirectly from this process. The examples also show that the MME functions tend to find local minima less often than LMS does for the same number of hidden units. This is believed to be due to the difference in network complexity needed to accurately approximate a posteriori probabilities versus class boundaries.
引用
收藏
页码:809 / 818
页数:10
相关论文
共 23 条
[1]   OPTIMIZATION FOR TRAINING NEURAL NETS [J].
BARNARD, E .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 1992, 3 (02) :232-240
[2]   PERFORMANCE AND GENERALIZATION OF THE CLASSIFICATION FIGURE OF MERIT CRITERION FUNCTION [J].
BARNARD, E .
IEEE TRANSACTIONS ON NEURAL NETWORKS, 1991, 2 (02) :322-325
[3]   A COMPARISON BETWEEN CRITERION FUNCTIONS FOR LINEAR CLASSIFIERS, WITH AN APPLICATION TO NEURAL NETS [J].
BARNARD, E ;
CASASENT, D .
IEEE TRANSACTIONS ON SYSTEMS MAN AND CYBERNETICS, 1989, 19 (05) :1030-1041
[4]   MINIMUM COMPLEXITY DENSITY-ESTIMATION [J].
BARRON, AR ;
COVER, TM .
IEEE TRANSACTIONS ON INFORMATION THEORY, 1991, 37 (04) :1034-1054
[5]  
Duda R. O., 1973, PATTERN CLASSIFICATI, V3
[6]  
Fletcher R., 1981, PRACTICAL METHODS OP
[7]  
FUKUNAGA K, 1990, INTRO STATISTICAL PA
[8]  
HAMPSHIRE J, 1991, 1990 P CONN MOD SUMM, P159
[9]  
HAMPSHIRE J, 1992, ADV NEURAL INFORMATI, V4
[10]  
HAMPSHIRE J, 1966, P SOC PHOTO-OPT INS, P76