摘要
大多数时间序列往往具有变方差的非线性特性,即某些时期的波动特别剧烈,而另一时期的波动又相对平 稳。对中国股票市场的非线性现象进行分析,发现上证综合指数和深证成分指数分布显示中国股票市场非线性现 象十分明显。在分析数据的基础上,建立了上证综合指数和深证成分指数的广义自回归条件异方差(GARCH)和 自回归移动平均(ARMA)预测模型,并分析了中国股票市场的几个非线性特征。
Most variances of time series have non-linear characteristics, that is, the variances change acutely sometimes, while they stay stable at other time. The non-linear feature of Chinese stock market is analyzed. The data distribution of Shanghai and Shenzhen stock markets shows that the non-linear feature of Chinese security market is obvious. GARCH and ARMA prediction models are constructed on condition of the analyzed data, and several conclusions about the non-linear characteristics of Chinese security market are drawn.
出处
《天津大学学报（社会科学版）》
2005年第6期417-420,共4页
Journal of Tianjin University：Social Sciences
关键词
收益率
广义自回归条件异方差
自回归移动平均
rate of return
generilized auto regressive conditional heteroskedastic
auto regressive moving average
