对称α稳定分布特性及其参数估计方法研究

Symmetrical Alpha Stable Distribution and Its Parameter Estimation Method

对称α稳定分布能较好地描述混响等非高斯对称概率密度序列特征,首先探讨并给出序列非高斯性、对称性判定方法,在此基础上,研究并比较分数低阶矩法和对数矩法两种对称α稳定分布参数估计方法,仿真及实测数据处理结果表明:对数矩法估计性能优于分数低阶矩法。

Symmetrical alpha stable distribution can describe sequence with symmetry and non-Gaussian probability density such as reverberation.The paper first discusses and gives the method of judging symmetry and non-Gaussian property of probability density.Then based on this,two parameter estimation methods for symmetrical alpha stable distribution that are fractional lower-order moment method and logarithmic moment method are studied and compared.Simulation and experimental results show that the logarithmic moment method is preferable to fractional lower-order moment method.