Abram M. Kagan Professor   Department of Mathematics University of Maryland College Park, MD 20742-4015 Room 1107 phone: 1-301-405-5061 fax: 1-301-314-0827 amk@math.umd.edu

Teaching

2010/2011

2011/2012

Research Areas

• Parameter Estimation  
• Fisher Information  
• Characterization Problems  
• Sufficiency and Exponential Families  
• Generalized Linear Models

Publications

• 2007
  i. Sub- and superadditivity a la Carlen of matrices related to the Fisher information (with Z. Landsman and C. R. Rao). J. Statist. Planning and Inference, 137, 291-298.
  ii. A lemma on stochastic majorization and properties of the Student distribution (with A. V. Nagaev). Theory Probab. Applications, 52, no. 1.
  iii. Strong decomposition of random variables (with J. Hoffman –Jorgensen, L. Pitt and L. Shepp) J. Theoret. Probab.
  iv. An identity for the Fisher information and Mahalanobis distance (with Bing Li) (submitted).
• 2006
  i. Quasi-independence of random variables and a property of normal and gamma distributions. J. Statist. Planning and Inference, 136, 199-208.
  ii. Profile sufficiency . Austrian J. Statistics, 35, 121-130.
• 2005
  i. A sufficiency paradox: an insufficient statistic preserving the Fisher information (with L. Shepp). American Statistician, 59, No. 1.
  ii. On estimation of a location parameter in presence of an ancillary component (with C. R. Rao). Theory Probab. and Its Applications, 50, 172-176.
  iii. The structure of the UMVUEs from categorical data (with M. Konikov). Theory Probab. and Its Applications, 50, No. 3.
  
• 2004
  i. On the maximum correlation coefficient (with W. Bryc and A. Dembo). Theory Probab. and Its Applications, 49, 191-197.
• 2003
  i. Some properties and applications of the efficient Fisher score (with C. R. Rao). J. Statist. Planning and Inference, 116, 343-352.
  ii. Statistical approach to some mathematical problems. Austrian J. of Statistics, 32, 71-83.
• 2002
  i. Sufficiency and ancillarity in characterization problems. J. Statist. Planning and Inference, 102, 223-228.
  ii. Transfer theorems in exponential families. Statist. and Probab. Letters, 57, 225-228.
  iii. An inequality for the Pitman estimators related to the Stam inequality. Sankhya .A64, 282-292
• 2001
  i. A discrete version of the Stam inequality and a characterization of the Poisson distribution. J. Statist. Planning and Inference, 92, 7-12.
  ii. A property of linear forms of independent random variables related to uniqueness of linear structure (with R.G. Laha). J. Statist. Planning and Inference, 92, 13-20.
  iii. Multivariate normal distributions, Fisher information and matrix inequalities (with P. J. Smith). Internat. J. of Mathem. Education in Science and Technology, 32, 91-96.
  iv. A note on the logistic link function. Biometrika, 88, 599-601.
  v. Remarks on the maximum correlation coefficient (with A. Dembo and L. Shepp). Bernoulli, 7, 343-350.
  vi. Another look at the Cramer-Rao inequality. American Statistician, 55, 211-212. vii. How many moments can be estimated from a large sample? (with S. Nagaev). Statist. and Probab. Letters, 55, 99-105.
• 2000
  i. An extension of Darmois-Skitovich theorem to a class of dependent random variables (with J. Wesolowski). Statist. and Probab. Letters, 47, 69-73
• 1999
  i. Symmetrization of binary random variables (with C. Mallows, L. Shepp, R. Vanderbei, and Y. Vardi). Bernoulli, 5, 1013-1020.
  ii. Relation between the covariance and Fisher information matrices (with Z. Landsman). Statist. and Probab. Letters, 42, 7-13.
  iii. Characterization of the Weibull distribution by properties of the Fisher information under type I censoring (with I. Gertsbakh). Statist. and Probab. Letters, 42, 99-105.
  iv. Characterization of the normal distribution through the power of a one-way ANOVA (with G. Letac). J. Statist. Planning and Inference, 77, 1-9.
  v. Tail hypotheses in the signal plus noise model (with L. Shepp). Statist. and Probab. Letters, 43, 317-319.
• 1998
  i. Some linear models are necessarily parametric (with L. Shepp). Statist. and Probab. Letters, 37, 77-80.
  ii. Why the variance? (with L Shepp). Statist. and Probab. Letters
  iii. A stronger version of matrix convexity as applied to functions of Hermitian matrices (with P.J. Smith). J. of Inequalities and Applications
• 1997
  i. Statistical meaning of Carlen's superadditivity of the Fisher information (with Z. Landsman). Statist. and Probab. Letters, 32, 175-179.
  ii. A class of polynomials with an invariance property (with V. Rohatgi). .J. Statist. Planning and Inference, 63, 215-222.
  iii. Independence of the sum and absolute difference of independent random variables does not imply their normality (with R. G. Laha and V. Rohatgi). Mathem. Methods of Statist., 6, 263-265.
• 1996
  i. Normality via conditional normality of linear forms (with J. Wesolowski). Statist. and Probab. Letters, 29, 229-232

In University

Faculty Advisor, University Chess Club, 1988 - president.

In Math Community

Associate Editor of "Mathematical Methods in Statistics"
Co-organizer of the 5th Lukacs Symposium in Probability and Statistics (Bowling Green, Spring 1995)
Co-organizer of Special Session of the Amer. Math. Soc. Regional Meeting (College Park, Spring 1997)

Other activities

Student's Corner