基于偏Laplace正态数据下位置、均值回归模型的参数估计  

Parameter Estimation of Location and Mean RegressionModel Based on Skew Laplace Normal Data

作　　者：郑桂芬 吴刘仓[1] 聂兴锋 ZHENG Guifen;WU Liucang;NIE Xingfeng(Faculty of Science,Kunming University of Science and Technology,Kunming 650093,China)

机构地区：[1]昆明理工大学理学院,云南昆明650093

出　　处：《应用数学》2020年第3期747-756,共10页Mathematica Applicata

基　　金：国家自然科学基金项目(11861041,11261025)。

摘　　要：参数估计是一种基本的统计推断形式,也是统计学的一个重要分支.在分析偏态数据时,我们比较关注数据的众数、中位数和均值,但是偏Laplace正态数据的众数和中位数难以精确求出,因此用位置参数来近似代替.故本文提出偏Laplace正态数据下位置和均值回归模型,并研究该模型的参数估计,模拟和实例研究结果表明本文提出的模型和方法是科学合理的.Parameter estimation is a fundamental form of statistical inference and an important branch of statistics.In the analysis of skewed data,we pay more attention to the mode,median and mean of the data,but the mode and median of the Skew Laplace normal distribution are difficult to accurately calculate,so it is approximate to substitute location parameters for them.Therefore,the location and mean regression model under Skew Laplace normal distribution is proposed,and the parameter estimation of the model is studied.The results of simulation and case study show that the proposed model and method are scientific and reasonable.

关 键 词：偏Laplace正态分布 位置回归模型 均值回归模型 EM算法 

分 类 号：O212.1[理学—概率论与数理统计]