摘要
探讨了混沌理论与前馈人工神经网络相结合进行非线性时间序列预测的基本原理，并对影响短波通信的两个参数：年太阳黑子数和天１０ｃｍ辐射强度，在ＳＶＤ滤波的基础上进行了预测。
In this paper, the principles of nonlinear time series prediction based on the combination of Chaos theory and FeedForward Neural Network are set forth. A noisefiltering algorithm based on SVD (Singular Value Decomposition) is discussed too. Two parameters in shortwave communication: Sunspot number and 10cm flux are predicted based on the algorithm.
出处
《电波科学学报》
EI
CSCD
1999年第2期172-177,共6页
Chinese Journal of Radio Science
关键词
混沌
非线性时间序列
前向BP神经网络
预测算法
Chaos
Nonlinear time series
Feedforward BP Neural Network
Sunspot Number
10cm flux
