上证综合指数的价格趋势预测模型研究

Research on Forecasting Model in Price Trend of Shanghai Composite lndex

越来越多的研究表明 ,股票价格序列有区别于随机游走所产生的序列。运用了Taylor的价格趋势模型及Josephine ,W .C .Kwanetal关于此模型的拟极大似然估计 (QML)方法估计参数过程中而得到的最小均方预测方程 ,对上证综合指数 (SCT)的年收益分区间进行建模与预测 ,得到了较好的预测效果。估计出的参数还表明 :(1)关于SCI的转折信号的方差大于随机波动方差 ;(2 )自 1994年以来 ,SCI涨跌趋势持续时间在 2天上下浮动。

More and more researches indicate that the stock price series is different from the series of the random walk.In this paper, we develop model and forecast annual price of Shanghai Composite Index(SCI) in sub-periods using the Taylor's price trend model and Josephine, W.C.et al the least square forecasting equation which is a by-products in estimating parameters of the model with quasi-maximum likelihood(QML). The estimated parameters also indicate that:(1)the variance of switching signal of SCI is much larger than the variance of random vaiance;(2)the duration time of the trend about SCI's rising and falling fluctuates around two days.