Volume 1,Issue 3
Fall 2025
流数据下的在线可更新稳健期望分位数回归
随着实际数据分析中动态流式数据集的比例不断上升，流数据的应用场景正在不断拓宽。在海量数据下，尖峰厚尾数
据分布占据重要比例，如何进行对应的稳健估计是非常有必要的。本研究提出基于Huber损失函数的稳健期望分位
数回归方法，仅使用历史汇总统计量实现在重尾噪声下给出实时高效的稳健估计，并且在特定假设条件下建立估计量
的渐近性质。模拟研究进一步验证，该方法在流式计算环境中处理大规模数据集时，具有稳健性和实时性。
[1]Luo L, Song P. Renewable estimation and incremental inference in generalized
linearmodels with streaming data sets[J]. Journal of the Royal Statistical Society:
Series B (Statistical Methodology), 2020,82(1), 69–97.
[2]Schifano E D, Wu J, Wang C, Yan J, Chen M H. Online updating of statistical
inference in the big data setting[J].Technometrics, 2016, 58(3):393-403.
[3]Wang K, Wang H, Li S. (2022). Renewable quantile regression for streaming
datasets. Knowledge-Based Systems, 235:107675.
[4]Huber P J. Robust estimation of a location parameter[J]. Annals of Mathematical
Statistics，1964,35 (1): 73-101.
[5]Gupta D, Hazarika B B, Berlin M. Robust regularized extreme learning machine
with asymmetric Huber loss function[J]. Neural Computing and Applications. 2020,32
(16): 12971–12998.
[6]Yi C, Huang J. Semismooth Newton coordinate descent algorithm for elastic-net
penalized Huber loss regression and quantile regression[J]. Journal of Computational
and Graphical Statistics, 2017, 26(3): 547-557.
[7]Sun Q, Zhou W X. Fan J. Adaptive Huber regression[J]. Journal of the American
Statistical Association, 2020, 115(529): 254-265.
[8] 俞搏天. p-Huber 损失函数及其鲁棒性研究 [D]; 浙江师范大学, 2021.
[9]Akkaya D, Pınar M Ç. Minimizers of sparsity regularized huber loss function[J].
Journal of Optimization Theory and Applications. 2020,187(1): 205-233.
[10]潘莹丽, 刘展, 朱千慧子. 大数据背景下基于Huber回归模型的分布式优化方法研
究[J]. 数理统计与管理, 2022, 41(04): 633-646.
[11]Man R, Tan K, Wang Z, Zhou W. Retire: Robust expectile regression in high
dimensions[J]. Journal of Econometrics, 2023,239 (2).
[12]Wang K, Wang H, Li S. Online renewable quantile regression and penalized
learning for streaming datasets[J]. IEEE Transactions on Neural Networks and
Learning Systems,2021,34(5),2341–2352.