Positive semidefiniteness of estimated covariance matrices in linear models for sample survey data

Stephen Haslett

Special Matrices (2016)

  • Volume: 4, Issue: 1, page 218-224
  • ISSN: 2300-7451

Abstract

top
Descriptive analysis of sample survey data estimates means, totals and their variances in a design framework. When analysis is extended to linear models, the standard design-based method for regression parameters includes inverse selection probabilities as weights, ignoring the joint selection probabilities. When joint selection probabilities are included to improve estimation, and the error covariance is not a diagonal matrix, the unbiased sample based estimator of the covariance is the Hadamard product of the population covariance, the elementwise inverse of selection probabilities and joint selection probabilities, and a sample selection matrix of rank equal to the sample size. This Hadamard product is however not always positive definite, which has implications for best linear unbiased estimation. Conditions under which a change in covariance structure leaves BLUEs and/or BLUPs are known. Interestingly, this class of “equivalent” matrices for linear models includes non-positive semi-definite matrices. The paper uses these results to explore how the estimated covariance from the sample can be modified so that it meets necessary conditions to be positive semidefinite, while retaining the property that fitting a linear model to the sampled data yields the same BLUEs and/or BLUPs as when the original Hadamard product is used.

How to cite

top

Stephen Haslett. "Positive semidefiniteness of estimated covariance matrices in linear models for sample survey data." Special Matrices 4.1 (2016): 218-224. <http://eudml.org/doc/277101>.

@article{StephenHaslett2016,
abstract = {Descriptive analysis of sample survey data estimates means, totals and their variances in a design framework. When analysis is extended to linear models, the standard design-based method for regression parameters includes inverse selection probabilities as weights, ignoring the joint selection probabilities. When joint selection probabilities are included to improve estimation, and the error covariance is not a diagonal matrix, the unbiased sample based estimator of the covariance is the Hadamard product of the population covariance, the elementwise inverse of selection probabilities and joint selection probabilities, and a sample selection matrix of rank equal to the sample size. This Hadamard product is however not always positive definite, which has implications for best linear unbiased estimation. Conditions under which a change in covariance structure leaves BLUEs and/or BLUPs are known. Interestingly, this class of “equivalent” matrices for linear models includes non-positive semi-definite matrices. The paper uses these results to explore how the estimated covariance from the sample can be modified so that it meets necessary conditions to be positive semidefinite, while retaining the property that fitting a linear model to the sampled data yields the same BLUEs and/or BLUPs as when the original Hadamard product is used.},
author = {Stephen Haslett},
journal = {Special Matrices},
language = {eng},
number = {1},
pages = {218-224},
title = {Positive semidefiniteness of estimated covariance matrices in linear models for sample survey data},
url = {http://eudml.org/doc/277101},
volume = {4},
year = {2016},
}

TY - JOUR
AU - Stephen Haslett
TI - Positive semidefiniteness of estimated covariance matrices in linear models for sample survey data
JO - Special Matrices
PY - 2016
VL - 4
IS - 1
SP - 218
EP - 224
AB - Descriptive analysis of sample survey data estimates means, totals and their variances in a design framework. When analysis is extended to linear models, the standard design-based method for regression parameters includes inverse selection probabilities as weights, ignoring the joint selection probabilities. When joint selection probabilities are included to improve estimation, and the error covariance is not a diagonal matrix, the unbiased sample based estimator of the covariance is the Hadamard product of the population covariance, the elementwise inverse of selection probabilities and joint selection probabilities, and a sample selection matrix of rank equal to the sample size. This Hadamard product is however not always positive definite, which has implications for best linear unbiased estimation. Conditions under which a change in covariance structure leaves BLUEs and/or BLUPs are known. Interestingly, this class of “equivalent” matrices for linear models includes non-positive semi-definite matrices. The paper uses these results to explore how the estimated covariance from the sample can be modified so that it meets necessary conditions to be positive semidefinite, while retaining the property that fitting a linear model to the sampled data yields the same BLUEs and/or BLUPs as when the original Hadamard product is used.
LA - eng
UR - http://eudml.org/doc/277101
ER -

References

top
  1. [1] D. Basu. An essay on the logical foundations of survey sampling, Part One, In Foundations of Statistical Inference, V.P. Godambe and D.A. Sprott (eds). Holt, Rinehart & Winston, 203-242, 1971.  
  2. [2] C-M. Cassel, C.E Särndal, J.H. Wretman. Foundations of Inference in Survey Sampling, Wiley, 1977.  Zbl0391.62007
  3. [3] R.L. Chambers, C.J. Skinner (eds.). Analysis of Survey Data, Wiley, 2003.  
  4. [4] W.G. Cochran. Sampling Techniques, Wiley, 1953.  
  5. [5] W.G. Cochran. Sampling Techniques, 3rd edition, Wiley, 1977.  Zbl0353.62011
  6. [6] S. Fisk. A very short proof of Cauchy’s interlace theorem for eigenvalues of Hermitian matrices, arXiv:math/0502408v1  
  7. [math.CA] 18 Feb 2005, http://arxiv.org/pdf/math/0502408v1.pdf, 2005.  
  8. [7] W.A. Fuller. Sampling Statistics, Wiley, 2009.  
  9. [8] M.H. Hansen, W.N. Hurwitz, W.G. Madow. Sample Survey Methods and Theory, Volumes 1 and 2, Wiley, 1953.  Zbl0052.14801
  10. [9] S. Haslett. The linear non-homogeneous estimator in sample surveys, Sankhyä Ser. B, 47(1985), 101-117.  Zbl0589.62008
  11. [10] S. Haslett, S. Puntanen. Equality of the BLUEs and/or BLUPs under two linear models using stochastic restrictions, Statistical Papers, 51(2010), 465-475. [WoS][Crossref] Zbl1247.62167
  12. [11] D.G. Horvitz, D.J. Thompson. A generalization of sampling without replacement from a finite universe, Journal of the American Statistical Association, 47(1952) 663-685. [Crossref] Zbl0047.38301
  13. [12] R. Lehtonen, E. Pahkinen. Practical Methods for Design and Analysis of Complex Surveys, 2nd Edition, Wiley, 2003.  Zbl0867.62004
  14. [13] A. Mercer, P. Mercer. Cauchy’s interlace theorem and lower bounds for the spectral radius, International Journal of Mathematics and Mathematical Sciences, 23(2000), 563-566.  Zbl0955.15007
  15. [14] C.R. Rao. A note on a previous lemma in the theory of least squares and some further results, Sankhyä Ser. A 30(1968), 259-266.  Zbl0197.15802
  16. [15] R.M. Royal, W.G.Cumberland. Variance estimation in finite population sampling, Journal of the American Statistical Association, 73(1978) 351-358. [Crossref] 
  17. [16] C-E. Särndal, B. Swensson, J. Wretman. Model Assisted Survey Sampling, Springer, 1992.  Zbl0742.62008
  18. [17] C.J. Skinner, D. Holt, T.M.F. Smith (eds.). Analysis of Complex Surveys, Wiley, 1989.  Zbl0701.62018

NotesEmbed ?

top

You must be logged in to post comments.

To embed these notes on your page include the following JavaScript code on your page where you want the notes to appear.

Only the controls for the widget will be shown in your chosen language. Notes will be shown in their authored language.

Tells the widget how many notes to show per page. You can cycle through additional notes using the next and previous controls.

    
                

Note: Best practice suggests putting the JavaScript code just before the closing </body> tag.