论文部分内容阅读
In this note we provide a simple framework to study Z-estimators in high dimensions.Of interest is the setting where the number of parameters p < n increases with the sample size n.As the main result,strong high dimensional linearization of Z-estimators proves to hold only under a classic Huber-type smoothness condition on the estimating equations.Our results apply to spatial median,giving sharp exponential-type inequalities.They extend the results in He and Shao(2000)and Zou et al.(2013).