拉普拉斯分布参数的近似贝叶斯估计

Approximate Bayesian Estimation of the Parameters of Laplace Distribution

拉普拉斯分布是刻画尖峰厚尾数据的重要分布之一.本文提出拉普拉斯分布两参数具有显式解的线性近似贝叶斯估计,通过理论证明和数值模拟验证了线性近似贝叶斯估计相比其他估计的优越性,并考察了线性近似贝叶斯估计随着样本量增加的渐近性质.

The Laplacian distribution is one of the most important distributions used to characterize the peak and thick-tailed data.This paper proposes a linear approximation Bayesian estimation with explicit solutions for the two parameters of the Laplace distribution.The superiority of linear approximate Bayesian estimation over other estimators is verified by theoretical derivation and numerical simulations,and the asymptotic behavior of the linear estimation with the increase of sample size is investigated.