摘要
文章结合经验似然统计推断技术,对带有均值漂移的高维线性模型,给出了一个基于经验似然的模型异常值检测方法,该方法允许模型参数的维数随样本量的增加而趋于无穷,数据模拟表明所提出的异常值检测方法是行之有效的。
Based on the empirical likelihood statistical inference technology, this paper presents a model outlier detection method based on empirical likelihood for high-dimensional linear models with mean drifting. This method allows the dimension of model parameters to tend to infinity with the increase of sample size. Data simulation shows that the outliers detection method is effective.
引文
[1]Hawkins D M, Bradu D, Kass G V. Location of Several Outliers in Multiple Regression Data Using Elemental Sets[J]. Technometrics,1984, 26(3).
[2]Breunig M M, Kriegel H P, Ng R T, et al. LOF:Identifying DensityBased Local Outliers[J]. ACM Sigmod Record, 2000, 29(2).
[3]Türkan S, Cetin M C, Toktamis O. Outlier Detection by Regression Diagnostics Based on Robust Parameter Estimates[J]. Hacettepe Journal of Mathematics and Statistics, 2012,(41).
[4]Cook R D. Detection of Influential Observation in Linear Regression[J]. Technometrics, 1977, 19(1).
[5]Owen A. Empirical Likelihood for Linear Models[J]. Annals of Statistics, 1991,(19).
[6]Owen A.Empirical Likelihood[M]. New York:Chapman and Hall,2001.
[7]石坚.高维线性模型中的经验似然[J].系统科学与数学, 2007, 27(1).
[8]Wang T,Li Z H. Outlier Detection in High-Dimensional Regression Model[J]. Communications in Statistics-Theory and Methods, 2016,46(14).