Constructing Confidence Regions of Optimal Expected Size

This article presents a Monte Carlo method for approximating the minimax expected size (MES) confidence set for a parameter known to lie in a compact set. The algorithm is motivated by problems in the physical sciences in which parameters are unknown physical constants related to the distribution of observable phenomena through complex numerical models. The method repeatedly draws parameters at random from the parameter space and simulates data as if each of those values were the true value of the parameter. Each set of simulated data is compared to the observed data using a likelihood ratio test. Inverting the likelihood ratio test minimizes the probability of including false values in the confidence region, which in turn minimizes the expected size of the confidence region. We prove that as the size of the simulations grows, this Monte Carlo confidence set estimator converges to the Γ-minimax procedure, where Γ is a polytope of priors. Fortran-90 implementations of the algorithm for both serial and parallel computers are available. We apply the method to an inference problem in cosmology.

[1]  Edward J. Wollack,et al.  Wilkinson Microwave Anisotropy Probe (WMAP) Three Year Results: Implications for Cosmology , 2006, astro-ph/0603449.

[2]  G. Chamberlain,et al.  Econometric applications of maxmin expected utility , 2000 .

[3]  A. Guth.  Inflationary universe: A possible solution to the horizon and flatness problems , 1981 .

[4]  E. Lehmann.  Testing Statistical Hypotheses , 1960 .

[5]  Kellen Petersen August.  Real Analysis , 2009 .

[6]  I. Dunsmore,et al.  Linear-loss interval estimation of location and scale parameters. , 1968, Biometrika.

[7]  M. Halpern,et al.  First-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Parameter Estimation Methodology , 2003 .

[8]  W. R. van Zwet,et al.  A Strong Law for Linear Functions of Order Statistics , 1980 .

[9]  E. L. Wright.  Constraints on Dark Energy from Supernovae, Gamma-Ray Bursts, Acoustic Oscillations, Nucleosynthesis, Large-Scale Structure, and the Hubble Constant , 2007, astro-ph/0701584.

[10]  Philip B. Stark,et al.  Using what we know: Inference with Physical Constraints , 2003 .

[11]  R. L. Winkler.  A Decision-Theoretic Approach to Interval Estimation , 1972 .

[12]  G. Casella,et al.  Evaluating Confidence Sets Using Loss Functions , 1989 .

[13]  Rainer Weiss,et al.  COBE Differential Microwave Radiometers - Instrument design and implementation , 1990 .

[14]  W. Nelson,et al.  Minimax Solution of Statistical Decision Problems by Iteration , 1966 .

[15]  Stefano Casertano,et al.  New Hubble Space Telescope Discoveries of Type Ia Supernovae at z ≥ 1: Narrowing Constraints on the Early Behavior of Dark Energy , 2006, astro-ph/0611572.

[16]  Christian P. Robert,et al.  Loss Functions for Set Estimations , 1994 .

[17]  M. Halpern,et al.  First-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: The Angular Power Spectrum , 2003, astro-ph/0302217.

[18]  G. Casella,et al.  Minimax Confidence Sets for the Mean of a Multivariate Normal Distribution , 1982 .

[19]  A. Cohen,et al.  Admissible Confidence Interval and Point Estimation for Translation or Scale Parameters , 1973 .

[20]  Peter J. Kempthorne,et al.  Numerical specification of discrete least favorable prior distributions , 1987 .

[21]  Samuel Karlin,et al.  Mathematical Methods and Theory in Games, Programming, and Economics , 1961 .

[22]  N. B. Suntzeff,et al.  Observational Constraints on the Nature of Dark Energy: First Cosmological Results from the ESSENCE Supernova Survey , 2007, astro-ph/0701041.

[23]  J. Pratt.  Length of Confidence Intervals , 1961 .

[24]  P. Stark,et al.  Minimax expected measure confidence sets for restricted location parameters , 2005 .

[25]  Patrick Billingsley,et al.  Probability and Measure. , 1986 .

[26]  J. R. Bond,et al.  Radical Compression of Cosmic Microwave Background Data , 2000 .

[27]  J. Robinson.  AN ITERATIVE METHOD OF SOLVING A GAME , 1951, Classics in Game Theory.

[28]  Edward J. Wollack,et al.  First year Wilkinson Microwave Anisotropy Probe (WMAP) observations: Determination of cosmological parameters , 2003, astro-ph/0302209.

[29]  Matemática,et al.  Society for Industrial and Applied Mathematics , 2010 .

[30]  Edward J. Wollack,et al.  First-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Preliminary Maps and Basic Results , 2003, astro-ph/0302207.

[31]  Jeff G. Schneider,et al.  Efficiently computing minimax expected-size confidence regions , 2007, ICML '07.

[32]  B. Vidakovic.  Γ-Minimax: A Paradigm for Conservative Robust Bayesians , 2000 .

[33]  J. Aitchison,et al.  Expected‐Cover and Linear‐Utility Tolerance Intervals , 1966 .

[34]  L. Berkovitz.  Convexity and Optimization in Rn , 2001 .

[35]  Ed Anderson,et al.  LAPACK Users' Guide , 1995 .

[36]  Peter M. Hooper.  Invariant Confidence Sets with Smallest Expected Measure , 1982 .

[37]  O. H. Brownlee,et al.  ACTIVITY ANALYSIS OF PRODUCTION AND ALLOCATION , 1952 .

[38]  Kenneth R. Davidson,et al.  Convexity and Optimization , 2009 .

[39]  A. Cohen,et al.  Admissibility Implications for Different Criteria in Confidence Estimation , 1973 .

[40]  U. Seljak,et al.  A Line of sight integration approach to cosmic microwave background anisotropies , 1996, astro-ph/9603033.

[41]  Edward J. Wollack,et al.  Three Year Wilkinson Microwave Anistropy Probe (WMAP) Observations: Polarization Analysis , 2006, astro-ph/0603450.

[42]  V. M. Joshi,et al.  Admissibility of the Usual Confidence Sets for the Mean of a Univariate or Bivariate Normal Population , 1969 .