L1-Penalized Quantile Regression in High Dimensional Sparse Models

We consider median regression and, more generally, quantile regression in high-dimensional sparse models. In these models the overall number of regressors p is very large, possibly larger than the sample size n, but only s of these regressors have non-zero impact on the conditional quantile of the response variable, where s grows slower than n. Since in this case the ordinary quantile regression is not consistent, we consider quantile regression penalized by the L1-norm of coefficients (L1-QR). First, we show that L1-QR is consistent at the rate of the square root of (s/n) log p, which is close to the oracle rate of the square root of (s/n), achievable when the minimal true model is known. The overall number of regressors p affects the rate only through the log p factor, thus allowing nearly exponential growth in the number of zero-impact regressors. The rate result holds under relatively weak conditions, requiring that s/n converges to zero at a super-logarithmic speed and that regularization parameter satisfies certain theoretical constraints. Second, we propose a pivotal, data-driven choice of the regularization parameter and show that it satisfies these theoretical constraints. Third, we show that L1-QR correctly selects the true minimal model as a valid submodel, when the non-zero coefficients of the true model are well separated from zero. We also show that the number of non-zero coefficients in L1-QR is of same stochastic order as s, the number of non-zero coefficients in the minimal true model. Fourth, we analyze the rate of convergence of a two-step estimator that applies ordinary quantile regression to the selected model. Fifth, we evaluate the performance of L1-QR in a Monte-Carlo experiment, and provide an application to the analysis of the international economic growth.

[1]  M. Loève On Almost Sure Convergence , 1951 .

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

[3]  H. Akaike A new look at the statistical model identification , 1974 .

[4]  T. C. Edens,et al.  Economic Growth , 1957, The Journal of Economic History.

[5]  G. Schwarz Estimating the Dimension of a Model , 1978 .

[6]  S. Geman A Limit Theorem for the Norm of Random Matrices , 1980 .

[7]  R. Levine,et al.  A Sensitivity Analysis of Cross-Country Growth Regressions , 1991 .

[8]  S. Portnoy Asymptotic behavior of regression quantiles in non-stationary, dependent cases , 1991 .

[9]  M. Talagrand,et al.  Probability in Banach Spaces: Isoperimetry and Processes , 1991 .

[10]  C. Gutenbrunner,et al.  Regression Rank Scores and Regression Quantiles , 1992 .

[11]  Moshe Buchinsky CHANGES IN THE U.S. WAGE STRUCTURE 1963-1987: APPLICATION OF QUANTILE REGRESSION , 1994 .

[12]  Yurii Nesterov,et al.  Interior-point polynomial algorithms in convex programming , 1994, Siam studies in applied mathematics.

[13]  P. Laplace Théorie analytique des probabilités , 1995 .

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

[15]  R. Tibshirani Regression Shrinkage and Selection via the Lasso , 1996 .

[16]  X. Sala-i-Martin,et al.  I Just Ran Two Million Regressions , 1997 .

[17]  R. Koenker,et al.  The Gaussian hare and the Laplacian tortoise: computability of squared-error versus absolute-error estimators , 1997 .

[18]  John N. Tsitsiklis,et al.  Introduction to linear optimization , 1997, Athena scientific optimization and computation series.

[19]  Keith Knight,et al.  Limiting distributions for $L\sb 1$ regression estimators under general conditions , 1998 .

[20]  R. Koenker,et al.  Goodness of Fit and Related Inference Processes for Quantile Regression , 1999 .

[21]  Wenjiang J. Fu,et al.  Asymptotics for lasso-type estimators , 2000 .

[22]  V. Chernozhukov Extremal quantile regression , 2005, math/0505639.

[23]  Jianqing Fan,et al.  Sure independence screening for ultrahigh dimensional feature space , 2006, math/0612857.

[24]  N. Meinshausen,et al.  High-dimensional graphs and variable selection with the Lasso , 2006, math/0608017.

[25]  Florentina Bunea,et al.  Aggregation and sparsity via 1 penalized least squares , 2006 .

[26]  Santosh S. Vempala,et al.  The geometry of logconcave functions and sampling algorithms , 2007, Random Struct. Algorithms.

[27]  A. Tsybakov,et al.  Aggregation for Gaussian regression , 2007, 0710.3654.

[28]  Terence Tao,et al.  The Dantzig selector: Statistical estimation when P is much larger than n , 2005, math/0506081.

[29]  S. Vempala,et al.  The geometry of logconcave functions and sampling algorithms , 2007 .

[30]  A. Belloni,et al.  On the Computational Complexity of MCMC-Based Estimators in Large Samples , 2007 .

[31]  R. Koenker,et al.  Regression Quantiles , 2007 .

[32]  A. Belloni,et al.  On the Computational Complexity of MCMC-Based Estimators in Large Samples , 2007, 0704.2167.

[33]  A. Tsybakov,et al.  Sparsity oracle inequalities for the Lasso , 2007, 0705.3308.

[34]  S. Geer HIGH-DIMENSIONAL GENERALIZED LINEAR MODELS AND THE LASSO , 2008, 0804.0703.

[35]  Cun-Hui Zhang,et al.  The sparsity and bias of the Lasso selection in high-dimensional linear regression , 2008, 0808.0967.

[36]  Aarnout Brombacher,et al.  Probability... , 2009, Qual. Reliab. Eng. Int..

[37]  N. Meinshausen,et al.  LASSO-TYPE RECOVERY OF SPARSE REPRESENTATIONS FOR HIGH-DIMENSIONAL DATA , 2008, 0806.0145.

[38]  Massimiliano Pontil,et al.  Taking Advantage of Sparsity in Multi-Task Learning , 2009, COLT.

[39]  A. Belloni,et al.  Least Squares After Model Selection in High-Dimensional Sparse Models , 2009, 1001.0188.

[40]  V. Koltchinskii Sparsity in penalized empirical risk minimization , 2009 .

[41]  P. Bickel,et al.  SIMULTANEOUS ANALYSIS OF LASSO AND DANTZIG SELECTOR , 2008, 0801.1095.

[42]  Pablo A. Parrilo,et al.  Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization , 2007, SIAM Rev..

[43]  A. Tsybakov,et al.  Sparse recovery under matrix uncertainty , 2008, 0812.2818.

[44]  H. Bateman Book Review: Ergebnisse der Mathematik und ihrer Grenzgebiete , 1933 .