HIERARCHICAL TRAINING OF NEURAL NETWORKS AND PREDICTION OF CHAOTIC TIME-SERIES

被引:18
作者
DEPPISCH, J [1 ]
BAUER, HU [1 ]
GEISEL, T [1 ]
机构
[1] UNIV FRANKFURT,SONDERFORSCHBEREICH NICHTLINEARE DYNAM,W-6000 FRANKFURT 11,GERMANY
关键词
D O I
10.1016/0375-9601(91)90340-E
中图分类号
O4 [物理学];
学科分类号
0702 ;
摘要
We present a new procedure for hierarchical training of multilayer perceptrons to outputs of high precision. It achieves a dramatic increase in accuracy, e.g. by three orders of magnitude, and can reduce training time considerably. The method is applied to the prediction of chaotic systems where we obtain the optimum error evolution for iterated predictions as well as a substantial reduction of the absolute prediction error.
引用
收藏
页码:57 / 62
页数:6
相关论文
共 21 条
[1]  
[Anonymous], 1987, LEARNING INTERNAL RE
[2]  
[Anonymous], 1988, EVOLUTION LEARNING C
[3]  
Bauer H.-U., 1989, International Journal of Neural Systems, V1, P187, DOI 10.1142/S0129065789000098
[4]   NONLINEAR DYNAMICS OF FEEDBACK MULTILAYER PERCEPTRONS [J].
BAUER, HU ;
GEISEL, T .
PHYSICAL REVIEW A, 1990, 42 (04) :2401-2409
[5]  
DEPPISCH J, 1990, THESIS U WURZBURG
[6]   PREDICTING CHAOTIC TIME-SERIES [J].
FARMER, JD ;
SIDOROWICH, JJ .
PHYSICAL REVIEW LETTERS, 1987, 59 (08) :845-848
[7]  
FLETCHER R, 1981, PRACTICAL METHODS OP, V1
[8]   ON DETERMINING THE DIMENSION OF CHAOTIC FLOWS [J].
FROEHLING, H ;
CRUTCHFIELD, JP ;
FARMER, D ;
PACKARD, NH ;
SHAW, R .
PHYSICA D, 1981, 3 (03) :605-617
[9]  
HAKIM NZ, 1990, P IJCNN90 SAN DIEGO, P593
[10]  
LAPEDES AS, 1987, LAUR87 LOS AL NAT LA