摘要
在概率统计中,偏度系数反映了随机变量的密度曲线的对称特征。由于偏度系数涉及到分布的前三阶矩,因此得到好的估计有一定的难度。文章建立贝叶斯模型,对偏度系数提出近似线性贝叶斯估计,并在多条数据结构下,对先验分布的超参数提出合适的估计,得到偏度系数的经验贝叶斯估计。
In probability and statistics, skewness coefficient reflects the symmetry characteristics of density curve of the ran- dom variable. As the skewness coefficient is involved in the first three moments of the distribution, it is very difficult to get good estimates for skewness coefficient. This paper builds Bayes model and proposes approximate linear Bayesian estimation for skew- ness coefficient. In addition, the hyper-parameters of prior distribution are estimated in the multitude data structure, thus deriving the empirical Bayes estimator of skewness coefficient.
出处
《统计与决策》
CSSCI
北大核心
2017年第10期78-81,共4页
Statistics & Decision
基金
国家自然科学基金资助项目(71361015)
教育部人文社科基金资助项目(15YJC910010)
江西省自然科学基金资助项目(20142BAB201013)
关键词
偏度系数
近似线性贝叶斯估计
超参数
经验贝叶斯估计
skewness coefficient, approximate linear Bayesian estimation
hyper-parameter
empirical Bayes estimation
