稳定分布的参数估计
The Estimation of the Parameter of Stable Distribution
-
摘要: 由于金融数据经常具有“高峰厚尾”现象,所以用稳定分布去拟合.但由于稳定分布没有密度函数显式,而且可能一阶矩或二阶矩又不存在,因此稳定分布的参数估计问题用经典方法很难处理.本文利用类似Duffie和Singleton(1993)的模拟矩法(SMM)的想法,构造了一种新的参数估计方法:并得到该估计的强相合性结果。最后举了一个实际的例子,分析了深圳成分指数的情况。Abstract: There are many distributions to fit the financial data, stable distribution is one kind of distribution classes which can better describe the financial data with heavy tail and excess kurtosis. It is very difficult to estimate the parameter of the stable distribution by the classical estimation methods for there is no direct density function and the first or higher moments may not exist. We develop a new parametric estimation similar as the Duffie and Singleton (1993)’s Simulation Moments Method (SMM) and discuss the property of the new estimator. Finally, we give a example of Shen Zhen composition index using our methods.