Adaptive Huber Regression

• 成果类型: Article
• 署名作者: Sun, Qiang; Zhou, Wen-Xin; Fan, Jianqing
• 署名单位: University of Toronto; University of California System; University of California San Diego; Fudan University; Princeton University
• 刊物名称: JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION
• ISSN/ISSBN: 0162-1459
• DOI: 10.1080/01621459.2018.1543124
• 发表日期: 2020
• 页码: 254-265
• 关键词: m-estimators; asymptotic-behavior; quantile regression; variable selection; robust regression; breast-cancer; data depth; expression; gene; parameters

摘要

Big data can easily be contaminated by outliers or contain variables with heavy-tailed distributions, which makes many conventional methods inadequate. To address this challenge, we propose the adaptive Huber regression for robust estimation and inference. The key observation is that the robustification parameter should adapt to the sample size, dimension and moments for optimal tradeoff between bias and robustness. Our theoretical framework deals with heavy-tailed distributions with bounded th moment for any . We establish a sharp phase transition for robust estimation of regression parameters in both low and high dimensions: when , the estimator admits a sub-Gaussian-type deviation bound without sub-Gaussian assumptions on the data, while only a slower rate is available in the regime and the transition is smooth and optimal. In addition, we extend the methodology to allow both heavy-tailed predictors and observation noise. Simulation studies lend further support to the theory. In a genetic study of cancer cell lines that exhibit heavy-tailedness, the proposed methods are shown to be more robust and predictive. for this article are available online.