Efficient Empirical Bayes prediction under check loss using Asymptotic Risk Estimates

We develop a novel Empirical Bayes methodology for prediction under check loss in high-dimensional Gaussian models. The check loss is a piecewise linear loss function having differential weights for measuring the amount of underestimation or overestimation. Prediction under it differs in fundamental aspects from estimation or prediction under weighted-quadratic losses. Because of the nature of this loss, our inferential target is a pre-chosen quantile of the predictive distribution rather than the mean of the predictive distribution. We develop a new method for constructing uniformly efficient asymptotic risk estimates which are then minimized to produce effective linear shrinkage predictive rules. In calculating the magnitude and direction of shrinkage, our proposed predictive rules incorporate the asymmetric nature of the loss function and are shown to be asymptotically optimal. Using numerical experiments we compare the performance of our method with traditional Empirical Bayes procedures and obtain encouraging results.

[1]  G. Szegö,et al.  Concerning sets of polynomials orthogonal simultaneously on several circles , 1939 .

[2]  E. L. Lehmann,et al.  Theory of point estimation , 1950 .

[3]  K. Arrow,et al.  Optimal Inventory Policy. , 1951 .

[4]  C. Stein Confidence Sets for the Mean of a Multivariate Normal Distribution , 1962 .

[5]  H. Robbins The Empirical Bayes Approach to Statistical Decision Problems , 1964 .

[6]  B. Efron,et al.  Combining Possibly Related Estimation Problems , 1973 .

[7]  B. Efron,et al.  Stein's Estimation Rule and Its Competitors- An Empirical Bayes Approach , 1973 .

[8]  B. Efron,et al.  Data Analysis Using Stein's Estimator and its Generalizations , 1975 .

[9]  Lawrence D. Brown,et al.  Estimation with Incompletely Specified Loss Functions (the Case of Several Location Parameters) , 1975 .

[10]  J. Berger Admissible Minimax Estimation of a Multivariate Normal Mean with Arbitrary Quadratic Loss , 1976 .

[11]  I. Good Some history of the hierarchical Bayesian methodology , 1980 .

[12]  C. Stein Estimation of the Mean of a Multivariate Normal Distribution , 1981 .

[13]  J. Aitchison,et al.  Statistical Prediction Analysis , 1975 .

[14]  C. Morris Parametric Empirical Bayes Inference: Theory and Applications , 1983 .

[15]  H. Robbins Asymptotically Subminimax Solutions of Compound Statistical Decision Problems , 1985 .

[16]  D. Dey,et al.  Estimation of a covariance matrix under Stein's loss , 1985 .

[17]  A. Zellner Bayesian Estimation and Prediction Using Asymmetric Loss Functions , 1986 .

[18]  W. Newey,et al.  Large sample estimation and hypothesis testing , 1986 .

[19]  Stephen M. Stigler,et al.  The 1988 Neyman Memorial Lecture: A Galtonian Perspective on Shrinkage Estimators , 1990 .

[20]  Edward I. George,et al.  Estimation under Profit-Driven Loss Functions , 1992 .

[21]  C. Stein,et al.  Estimation with Quadratic Loss , 1992 .

[22]  S. Thangavelu Lectures on Hermite and Laguerre expansions , 1993 .

[23]  I. Johnstone,et al.  Adapting to Unknown Smoothness via Wavelet Shrinkage , 1995 .

[24]  Bimal K. Sinha,et al.  A new general interpretation of the Stein estimate and how it adapts: Applications , 1999 .

[25]  Evan L. Porteus,et al.  Stalking Information: Bayesian Inventory Management with Unobserved Lost Sales , 1999 .

[26]  Kurt Hoffmann,et al.  Stein estimation—A review , 2000 .

[27]  Cun-Hui Zhang,et al.  Compound decision theory and empirical bayes methods , 2003 .

[28]  S. James Press,et al.  Subjective and objective Bayesian statistics : principles, models, and applications , 2003 .

[29]  I. Krasikov New bounds on the Hermite polynomials , 2004, math/0401310.

[30]  Richard L. Smith,et al.  PREDICTIVE INFERENCE , 2004 .

[31]  Herbert E. Scarf,et al.  Inventory Models of the Arrow-Harris-Marschak Type with Time Lag , 2005 .

[32]  R. Koenker Quantile Regression: Fundamentals of Quantile Regression , 2005 .

[33]  Feng Liang,et al.  Improved minimax predictive densities under Kullback-Leibler loss , 2006 .

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

[35]  L. Brown In-season prediction of batting averages: A field test of empirical Bayes and Bayes methodologies , 2008, 0803.3697.

[36]  Ya'acov Ritov,et al.  Asymptotic e-ciency of simple decisions for the compound decision problem ⁄ , 2008, 0802.1319.

[37]  T. Cai,et al.  Testing composite hypotheses, Hermite polynomials and optimal estimation of a nonsmooth functional , 2011, 1105.3039.

[38]  Ingo Steinwart,et al.  Estimating conditional quantiles with the help of the pinball loss , 2011, 1102.2101.

[39]  E. George,et al.  A Tribute to Charles Stein , 2012, 1203.4664.

[40]  Lawrence D. Brown,et al.  SURE Estimates for a Heteroscedastic Hierarchical Model , 2012, Journal of the American Statistical Association.

[41]  Gourab Mukherjee,et al.  EXACT MINIMAX ESTIMATION OF THE PREDICTIVE DENSITY IN SPARSE GAUSSIAN MODELS. , 2015, Annals of statistics.

[42]  Georgia Perakis,et al.  The Data-Driven Newsvendor Problem: New Bounds and Insights , 2015, Oper. Res..

[43]  L. Brown,et al.  OPTIMAL SHRINKAGE ESTIMATION OF MEAN PARAMETERS IN FAMILY OF DISTRIBUTIONS WITH QUADRATIC VARIANCE. , 2016, Annals of statistics.