Accommodation of outliers by robust MML estimation for spatial autoregressive model

被引:0
|
作者
Shukla, Sweta [1 ]
Lalitha, S. [2 ]
Srivastava, Pulkit [3 ]
机构
[1] GLA Univ, Inst Appl Sci & Humanities, Dept Math, Mathura 281406, Uttar Pradesh, India
[2] Univ Allahabad, Dept Stat, Prayagraj 211002, Uttar Pradesh, India
[3] Univ Delhi, Fac Math Sci, Dept Stat, Delhi 110007, India
关键词
MML estimation; Spatial autoregressive model; Outlier; Robust; Type II censoring; CENSORED NORMAL SAMPLES;
D O I
10.1007/s13198-023-01856-w
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
Some outliers might be undetected even after using the outlier detection procedures. In this paper, an accommodation procedure for such undetected outliers is discussed, which is done by robust modified maximum likelihood (MML) estimation of Type II censored sample for a Spatial Autoregressive (SAR) model. A new method is proposed for determining the number of observation to be censored to obtain robust parameter estimates. Next, a Monte-Carlo simulation study is carried to assess the robustness of the obtained MML estimators for both the normal and the contaminated normal models. Also, asymptotic variances, covariances, and distributional results are obtained for the MML estimators.
引用
收藏
页码:293 / 306
页数:14
相关论文
共 50 条