基于LAD的对称alpha稳定分布参数估计

Parameters Estimation of Symmetric alpha Stable Distribution Based on Least Absolute Deviation

稳定分布是一种典型的非高斯分布,常用来描述具有尖峰厚尾特点的数据。由于alpha稳定分布的特性,参数估计成为一个难点。在样本特征函数的基础上,探讨了用最小绝对偏差法估计对称alpha稳定分布的参数,并在大样本条件下给出了参数的渐近置信区间。

As a typical non-Gaussian distribution, alpha stable distribution is usually used to present data with heavy tail and excess kurtosis. Due to the characteristics of alpha stable distribution, parameters estimation is a difficulty. On the base of sample characteristic function, parameters estimation of symmetric alpha stable distribution is discussed by least absolute deviation method in this paper. Under the condition of large sample, asymptotic confidence intervals of parameters is given.