【摘要】 传统的时间序列预测模型在处理具有非线性特性或非平稳时间序列问题,特别是对有人参与的主动系统、社会经济系统的预测上,无法取得满意的预测效果.寻求处理这类系统的方法是人们一直努力的方向.这里以小波理论为基础,重点研究了小波网络在非线性时间序列中的建模预测方法,利用深圳综合指数数据,建立了股票指数预测模型.该模型克服了传统的时间序列预测模型仅局限于线性系统的情况,避免了BP神经网络模型固有的缺陷.仿真结果表明,该方法比神经网络预测方法的预测精度高,可以很好地应用于某些非线性时间序列的预测中.更多还原

【Abstract】 The traditional time series prediction model is not able to achieve a satisfying prediction effect in the problem of a non-linear system and nonstationary time series, especially the system in which people participate in or a social economic system. To solve these problems, this paper presents essential research on modeling and prediction with wavelet neural network in the non-linear time series. Using Shenzhen’s integrative index dates, a prediction model of the stock index is established. The model abstains from the default of traditional time series prediction model that only can be used in linear system; the intrinsic defects of BP neural network are avoided. The simulation results indicated that this method is more accurate than neural network prediction. Furthermore,this prediction model can be effecticely used in some non-linear time series prediction.更多还原