摘要
本文将股票价格走势作为一个混沌系统，利用混沌时间序列理论对股票价格走势进行分析。通过混沌时序相空间重构技术将股票价格走势对应到一个高维系统，以获得一个在一定区间不断重叠的运行轨迹。建立一个神经网络以拟舍重构相空间中轨迹。实验表明，无论在预测精度还是速度上，混沌时间序列和弹性反馈算法的结合，获得比一般神经网络预测更好的效果。
A new method is given to predict the movements of stock price as a chaotic system using the theory of chaotic time series in this paper.The movements of stock price are corresponded to a high-dimensional system to obtain a constantly overlapping trajectory within a certain range using phase space reconstruction technology.After that,a neural network can be established to fit the trajectories of the reconstructed phase space.Experiments show that,the cooperation of chaotic time series and resilient BP obtains better effect than general network in terms of prediction accuracy or speed.
出处
《计算机光盘软件与应用》
2011年第19期69-69,76,共2页
Computer CD Software and Application
关键词
时间序列
弹性BP
股票预测
神经网络
混沌
Time series
Resilient BP
Stock prediction
Neural network
Chaotic
