论文部分内容阅读
应用向量值状态空间模型的两步建模方法分析了深圳五种股票价格的动态行为，样本选在１９９６年下半年这一股市波动较为剧烈的时期．研究结果表明：在趋势模型中五种股票的价格序列构成了带有一个共同趋势（根的估计值为０．９９）的协调积分过程．此外在周期模型中还发现了一个较弱的趋势和两个复数根．最后通过有关统计量和２０天的样本外预测以及Ｈｅｎｒｉｋｓｓｏｎ与Ｍｅｒｔｏｎ提出的非参数检验方法验正了分析和预测结果的有效性
This paper analyzes the dynamic behavior of five stock prices in Shenzhen using the two-step modeling method of state space model of vector value. The sample is selected in the period when the stock market fluctuated more intense in the second half of 1996. The results show that the price series of the five stocks in the trend model form a coordinated integral process with a common trend (root’s estimated value of 0.99). In addition, we find a weak trend and two complex roots in the periodic model. Finally, the validity of the analysis and prediction results are verified by the relevant statistics and the 20-day out-of-sample prediction and the nonparametric test method proposed by Henriksson and Merton