报告时间：2018年12月26日 14:00—15:00
报告地点：明德主楼1016会议室
报告题目: High-Dimensional inference: Debiasing the debiased Lasso with bootstrap
报告摘要:
In this talk, we consider the problem of constructing confidence intervals for low-dimensional coefficients in high-dimensional linear regression models. We propose to further debias the debiased Lasso estimator with bootstrap and prove its consistency for distribution approximation under proper conditions. The proposed method admits weaker sample size conditions in existence of a large proportion of large coefficients and reveals the benefits of having strong signals.
报告人简介：