Adaptive Capped Least Squares

This paper proposes the capped least squares regression with an adaptive resistance parameter, hence the name, adaptive capped least squares regression. The key observation is, by taking the resistant parameter to be data dependent, the proposed estimator achieves full asymptotic efficiency without losing the resistance property: it achieves the maximum breakdown point asymptotically. Computationally, we formulate the proposed regression problem as a quadratic mixed integer programming problem, which becomes computationally expensive when the sample size gets large. The data-dependent resistant parameter, however, makes the loss function more convex-like for larger-scale problems. This makes a fast randomly initialized gradient descent algorithm possible for global optimization. Numerical examples indicate the superiority of the proposed estimator compared with classical methods. Three data applications to cancer cell lines, stationary background recovery in video surveillance, and blind image inpainting showcase its broad applicability.

[1]  V. Yohai,et al.  High Breakdown-Point Estimates of Regression by Means of the Minimization of an Efficient Scale , 1988 .

[2]  F. Hampel.  A General Qualitative Definition of Robustness , 1971 .

[3]  Bruce Western,et al.  Concepts and Suggestions for Robust Regression Analysis , 1995 .

[4]  Qiang Sun,et al.  Adaptive Huber Regression , 2017, Journal of the American Statistical Association.

[5]  Xiaohua Zhang,et al.  MAL2 promotes proliferation, migration, and invasion through regulating epithelial-mesenchymal transition in breast cancer cell lines. , 2018, Biochemical and biophysical research communications.

[6]  Rafael A. Irizarry,et al.  A Model-Based Background Adjustment for Oligonucleotide Expression Arrays , 2004 .

[7]  Kengo Kato,et al.  Gaussian approximation of suprema of empirical processes , 2012, 1212.6885.

[8]  Frederick R. Forst,et al.  On robust estimation of the location parameter , 1980 .

[9]  V. Yohai,et al.  A class of robust and fully efficient regression estimators , 2002 .

[10]  Christine H. Müller,et al.  Breakdown points for designed experiments , 1995 .

[11]  Jon A. Wellner,et al.  Weak Convergence and Empirical Processes: With Applications to Statistics , 1996 .

[12]  C. Niu,et al.  Downregulation of annexin A3 inhibits tumor metastasis and decreases drug resistance in breast cancer , 2018, Cell Death & Disease.

[13]  Runze Li,et al.  A Tuning-free Robust and Efficient Approach to High-dimensional Regression , 2020, Journal of the American Statistical Association.

[14]  Yanqiu Zhao,et al.  TISSUE-SPECIFIC STEM CELLS Inhibition of TGF-b/Smad Signaling by BAMBI Blocks Differentiation of Human Mesenchymal Stem Cells to Carcinoma-Associated Fibroblasts and Abolishes Their Protumor Effects , 2012 .

[15]  Jianqing Fan,et al.  Estimation of high dimensional mean regression in the absence of symmetry and light tail assumptions , 2017, Journal of the Royal Statistical Society. Series B, Statistical methodology.

[16]  V. Yohai.  HIGH BREAKDOWN-POINT AND HIGH EFFICIENCY ROBUST ESTIMATES FOR REGRESSION , 1987 .

[17]  P. J. Huber.  Robust Regression: Asymptotics, Conjectures and Monte Carlo , 1973 .

[18]  D. G. Simpson,et al.  Minimum Hellinger Distance Estimation for the Analysis of Count Data , 1987 .

[19]  Howard D Bondell,et al.  Efficient Robust Regression via Two-Stage Generalized Empirical Likelihood , 2013, Journal of the American Statistical Association.