带偏广义误差分布（SGED分布）及其应用研究

发布时间：2018-10-22 10:21

【摘要】：由于金融市场价格收益率的分布一般具有波动性聚集以及尖峰厚尾等特征,对其进行分析建模时就需要根据数据的特征来选择恰当的模型和适当的分布假设,而常规的分布,如正态分布等无法很好地描绘这一数据特征。为了能够更好地拟合此类数据,我们提出了一种新的带偏广义误差分布(Skewed Generalized Error Distribution),也就是SGED分布,这个分布具有长尾带偏的特性,并推导出了这个分布的一些统计性质,如密度分布函数、累积分布函数、分位数函数、数字特征值(均值、方差、峰度、偏度等)、信息矩阵、似然函数的一致性以及渐近正态性等。将该分布与Probit模型相结合,实证分析了国际黄金市场黄金价格的影响因素以及投资操作行为,结果表明SGED分布能够更好地描述国际黄金价格的波动特征,并且具有较强的预测能力。
[Abstract]:Since the distribution of price returns in financial markets generally has the characteristics of volatility aggregation and sharp peaks and thick tails, it is necessary to select the appropriate model and appropriate distribution assumptions according to the characteristics of the data when analyzing and modeling them. Such as normal distribution and so on can not well describe the characteristics of the data. In order to fit this kind of data better, we propose a new generalized error distribution (Skewed Generalized Error Distribution), that is, SGED distribution, which has the characteristics of long tail band bias, and derive some statistical properties of this distribution. Such as density distribution function, cumulative distribution function, quantile function, numerical eigenvalue (mean, variance, kurtosis, bias, etc.), information matrix, consistency of likelihood function and asymptotic normality. Combining this distribution with Probit model, the paper empirically analyzes the influencing factors of gold price and investment operation behavior in international gold market. The results show that SGED distribution can better describe the fluctuating characteristics of international gold price. And has a strong ability to predict.
【学位授予单位】：华东理工大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：O211.3

【共引文献】