广义矩估计的延伸——广义经验似然估计
发布时间:2018-03-09 04:21
本文选题:广义矩估计 切入点:广义经验似然类估计量 出处:《统计与信息论坛》2012年03期 论文类型:期刊论文
【摘要】:从广义矩估计(GMM)到广义经验似然估计(GEL)的发展,是由于GMM估计量小样本性质的不足,促使人们寻求方法的改进和拓展。通过必要的证明和推导,详细解析GEL类估计量(包括EL,ET,CUE)的逻辑关系和数理结构,认识GEL的内在本质,并运用随机模拟方法证实了在小样本场合GEL类估计量比GMM估计量具有更小的估计偏差和均方误差,即GEL类估计改进了GMM估计的小样本性质。
[Abstract]:The development from generalized moment estimation (GMM) to generalized empirical likelihood estimator (Gel) is due to the shortage of small sample properties of GMM estimator, which urges people to seek improvement and extension of the method. The logic relation and mathematical structure of GEL class estimator (including ELT et CUE) are analyzed in detail, and the intrinsic nature of GEL is recognized. The stochastic simulation method is used to prove that GEL class estimator has smaller estimation deviation and mean square error than GMM estimator in small sample cases. That is, GEL class estimators improve the small sample properties of GMM estimators.
【作者单位】: 西南财经大学统计学院;
【基金】:教育部社科研究基金西部和边疆地区项目《证券市场动态相关性测度的拓展及应用研究》(11XJC910001)
【分类号】:F064.1
,
本文编号:1586967
本文链接:https://www.wllwen.com/jingjilunwen/jjsxs/1586967.html