A POLYNOMIAL-TIME ALGORITHM FOR THE CONSTRUCTION AND TRAINING OF A CLASS OF MULTILAYER PERCEPTRONS

被引:54
作者
ROY, A
KIM, LS
MUKHOPADHYAY, S
机构
[1] Arizona State Univ, Tempe, United States
基金
美国国家科学基金会;
关键词
POLYNOMIAL TIME ALGORITHM; MULTILAYER PERCEPTRONS; LINEAR PROGRAMMING; CLASSIFICATION ALGORITHM; SUPERVISED LEARNING; CLUSTERING; NET DESIGN;
D O I
10.1016/S0893-6080(05)80057-7
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 a polynomial time algorithm for the construction and training of a class of multilayer perceptrons for classification. It uses linear programming models to incrementally generate the hidden layer in a restricted higher-order perceptron. Polynomial time complexity of the method is proven. Computational results are provided for several well-known applications in the areas of speech recognition, medical diagnosis, and target detection. In all cases, very small nets were created that had error rates similar to those reported so far.
引用
收藏
页码:535 / 545
页数:11
相关论文
共 29 条
[1]  
[Anonymous], 2003, LINEAR PROGRAMMING
[2]   ART-2 - SELF-ORGANIZATION OF STABLE CATEGORY RECOGNITION CODES FOR ANALOG INPUT PATTERNS [J].
CARPENTER, GA ;
GROSSBERG, S .
APPLIED OPTICS, 1987, 26 (23) :4919-4930
[3]  
Duda R. O., 1973, PATTERN CLASSIFICATI, V3
[4]  
Everitt B., 1980, CLUSTER ANAL
[5]  
FARLOW S, 1984, SELFORGANIZING METHO
[6]   EVALUATING ALTERNATIVE LINEAR-PROGRAMMING MODELS TO SOLVE THE 2-GROUP DISCRIMINANT PROBLEM [J].
FREED, N ;
GLOVER, F .
DECISION SCIENCES, 1986, 17 (02) :151-162
[7]  
Freed N., 1981, Decision Sciences, V12, P68, DOI 10.1111/j.1540-5915.1981.tb00061.x
[8]   LEARNING, INVARIANCE, AND GENERALIZATION IN HIGH-ORDER NEURAL NETWORKS [J].
GILES, CL ;
MAXWELL, T .
APPLIED OPTICS, 1987, 26 (23) :4972-4978
[9]   A NEW CLASS OF MODELS FOR THE DISCRIMINANT PROBLEM [J].
GLOVER, F ;
KEENE, S ;
DUEA, B .
DECISION SCIENCES, 1988, 19 (02) :269-280
[10]   ANALYSIS OF HIDDEN UNITS IN A LAYERED NETWORK TRAINED TO CLASSIFY SONAR TARGETS [J].
GORMAN, RP ;
SEJNOWSKI, TJ .
NEURAL NETWORKS, 1988, 1 (01) :75-89