The Cramér-Rao bound for estimation-after-selection

In many practical parameter estimation problems, a model selection is made prior to estimation. In this paper, we consider the problem of estimating an unknown parameter of a selected population, where the population is chosen from a population set by using a predetermined selection rule. Since the selection step may have an important impact on subsequent estimation, ignoring it could lead to biased-estimation and an invalid Cramér-Rao bound (CRB). In this work, the mean-square-selected-error (MSSE) criterion is used as a performance measure. The concept of Ψ-unbiasedness is introduced for a given selection rule, Ψ, by using the Lehmann-unbiasedness definition. We derive a non-Bayesian Cramér-Rao-type bound on the MSSE of any Ψ-unbiased estimator. The proposed Ψ-CRB is a function of the conditional Fisher information and is a valid bound on the MSSE. Finally, we examine the Ψ-CRB for different selection rules for mean estimation in a linear Gaussian model.

[1]  Hado van Hasselt,et al.  Estimating the Maximum Expected Value: An Analysis of (Nested) Cross Validation and the Maximum Sample Average , 2013, ArXiv.

[2]  Yonina C. Eldar,et al.  Rethinking biased estimation [Lecture Notes] , 2008, IEEE Signal Processing Magazine.

[3]  A. Parsian,et al.  Estimation After Selection Under Reflected Normal Loss Function , 2012 .

[4]  Alfred O. Hero,et al.  Optimal simultaneous detection and estimation under a false alarm constraint , 1995, IEEE Trans. Inf. Theory.

[5]  Karin Baier.  Selecting And Ordering Populations A New Statistical Methodology , 2016 .

[6]  Jack Bowden,et al.  Unbiased Estimation of Selected Treatment Means in Two‐Stage Trials , 2008, Biometrical journal. Biometrische Zeitschrift.

[7]  Lang Tong,et al.  On Topology Attack of a Smart Grid: Undetectable Attacks and Countermeasures , 2013, IEEE Journal on Selected Areas in Communications.

[8]  Sailes K. Sengijpta.  Fundamentals of Statistical Signal Processing: Estimation Theory , 1995 .

[9]  A. Neath.  Testing Statistical Hypotheses (3rd ed.). E. L. Lehmann and Joseph P. Romano , 2006 .

[10]  Eric Chaumette,et al.  Cramér-Rao Bound Conditioned by the Energy Detector , 2007, IEEE Signal Processing Letters.

[11]  Yonina C. Eldar,et al.  Rethinking Biased Estimation , 2008 .

[12]  P. Stoica,et al.  On the Cramer-Rao bound for model-based spectral analysis , 2002, IEEE Signal Processing Letters.

[13]  Joseph Tabrikian,et al.  Non-Bayesian Periodic Cramér-Rao Bound , 2013, IEEE Transactions on Signal Processing.

[14]  Joseph Tabrikian,et al.  Performance bounds for constrained parameter estimation , 2012, 2012 IEEE 7th Sensor Array and Multichannel Signal Processing Workshop (SAM).

[15]  Alfred O. Hero,et al.  Optimal Two-Stage Search for Sparse Targets Using Convex Criteria , 2008, IEEE Transactions on Signal Processing.

[16]  H. Vincent Poor,et al.  Sensor Selection in Distributed Multiple-Radar Architectures for Localization: A Knapsack Problem Formulation , 2012, IEEE Transactions on Signal Processing.

[17]  P. Larzabal,et al.  On the influence of a detection step on lower bounds for deterministic parameter estimation , 2005, IEEE Transactions on Signal Processing.

[18]  E. Handschin,et al.  Bad data analysis for power system state estimation , 1975, IEEE Transactions on Power Apparatus and Systems.

[19]  Stephen E. Fienberg,et al.  Testing Statistical Hypotheses , 2005 .

[20]  Boaz Nadler,et al.  Parametric joint detection-estimation of the number of sources in array processing , 2010, 2010 IEEE Sensor Array and Multichannel Signal Processing Workshop.

[21]  Alfred O. Hero,et al.  Two-Stage Multiscale Search for Sparse Targets , 2011, IEEE Transactions on Signal Processing.

[22]  Tong Zhao,et al.  Distributed Sequential Bayesian Estimation of a Diffusive Source in Wireless Sensor Networks , 2007, IEEE Transactions on Signal Processing.

[23]  Harold B. Sackrowitz,et al.  A Decision Theory Formulation for Population Selection Followed by Estimating the Mean of the Selected Population , 1988 .

[24]  M. L. Eaton.  Some Optimum Properties of Ranking Procedures , 1967 .