Global statistical information in exponential experiments and selection of exponential models

Igor Vajda; E. van der Meulen

Applications of Mathematics (1998)

  • Volume: 43, Issue: 1, page 23-51
  • ISSN: 0862-7940

Abstract

top
The concept of global statistical information in the classical statistical experiment with independent exponentially distributed samples is investigated. Explicit formulas are evaluated for common exponential families. It is shown that the generalized likelihood ratio test procedure of model selection can be replaced by a generalized information procedure. Simulations in a classical regression model are used to compare this procedure with that based on the Akaike criterion.

How to cite

top

Vajda, Igor, and Meulen, E. van der. "Global statistical information in exponential experiments and selection of exponential models." Applications of Mathematics 43.1 (1998): 23-51. <http://eudml.org/doc/32995>.

@article{Vajda1998,
abstract = {The concept of global statistical information in the classical statistical experiment with independent exponentially distributed samples is investigated. Explicit formulas are evaluated for common exponential families. It is shown that the generalized likelihood ratio test procedure of model selection can be replaced by a generalized information procedure. Simulations in a classical regression model are used to compare this procedure with that based on the Akaike criterion.},
author = {Vajda, Igor, Meulen, E. van der},
journal = {Applications of Mathematics},
keywords = {exponential families; information divergence; Fisher information; global information; Akaike criterion; model selection; global information; exponential families; information divergence; Akaike criterion},
language = {eng},
number = {1},
pages = {23-51},
publisher = {Institute of Mathematics, Academy of Sciences of the Czech Republic},
title = {Global statistical information in exponential experiments and selection of exponential models},
url = {http://eudml.org/doc/32995},
volume = {43},
year = {1998},
}

TY - JOUR
AU - Vajda, Igor
AU - Meulen, E. van der
TI - Global statistical information in exponential experiments and selection of exponential models
JO - Applications of Mathematics
PY - 1998
PB - Institute of Mathematics, Academy of Sciences of the Czech Republic
VL - 43
IS - 1
SP - 23
EP - 51
AB - The concept of global statistical information in the classical statistical experiment with independent exponentially distributed samples is investigated. Explicit formulas are evaluated for common exponential families. It is shown that the generalized likelihood ratio test procedure of model selection can be replaced by a generalized information procedure. Simulations in a classical regression model are used to compare this procedure with that based on the Akaike criterion.
LA - eng
KW - exponential families; information divergence; Fisher information; global information; Akaike criterion; model selection; global information; exponential families; information divergence; Akaike criterion
UR - http://eudml.org/doc/32995
ER -

References

top
  1. Information theory and an extension of the maximum likelihood principle, Proceedings of the Second International Symposium on Information Theory, B. N. Petrov et al. (eds.), Akademiai Kiado, Budapest, 1973, pp. 267–281. (1973) Zbl0283.62006MR0483125
  2. 10.1080/02331888808802068, Statistics 19 (1988), 39–44. (1988) MR0921623DOI10.1080/02331888808802068
  3. Identification of a univariate ARMA model, Comp. Statist. 9 (1994), 117–133. (1994) MR1280755
  4. Fundamentals of Statistical Exponential Families, Inst. of Mathem. Statist, Hayword, California, 1986. (1986) Zbl0685.62002MR0882001
  5. 10.1214/aoms/1177728725, Ann. Math. Statist. 25 (1954), 573–578. (1954) MR0065087DOI10.1214/aoms/1177728725
  6. Information Theory and Statistics, Wiley, New York, 1959. (1959) Zbl0088.10406MR0103557
  7. Testing Statistical Hypotheses, 2nd edition, Wiley, New York, 1986. (1986) Zbl0608.62020MR0852406
  8. Convex Statistical Distances, Teubner, Leipzing, 1987. (1987) MR0926905
  9. Regression and Autoregression Models of Signals and their Recognition by Neural Nets, Diploma Theses. Faculty of Physical and Nuclear Engineering, Czech Tech. University, Prague, 1996. (Czech) (1996) 
  10. 10.1016/0047-259X(88)90137-6, Journal of Multivariate Analysis 27 (1988), 392–403. (1988) Zbl0684.62026MR0970962DOI10.1016/0047-259X(88)90137-6
  11. 10.1214/aos/1176346253, Ann. Statist. 11 (1983), 872–885. (1983) MR0707937DOI10.1214/aos/1176346253
  12. 10.1016/0005-1098(78)90005-5, Automatica 14 (1978), 465–471. (1978) DOI10.1016/0005-1098(78)90005-5
  13. Convex Analysis, Princeton University Press, Princeton, 1970. (1970) 
  14. Robust model selection, In Transactions of the Twelfth Prague Conference on Information Theory, ..., J. Á. Víšek, and P. Lachout (eds.), Academy of Sciences of the Czech Republic, Prague, 1994, pp. 200–202. (1994) 
  15. 10.1080/02331889508802501, Statistics 26 (1995), 345–354. (1995) MR1365683DOI10.1080/02331889508802501
  16. Akaike Information Criterion Statistics, Reidel, Dordrecht, 1986. (1986) 
  17. 10.1214/aos/1176344136, Annals of Statistics 6 (1978), 461–464. (1978) MR0468014DOI10.1214/aos/1176344136
  18. An Atlas of Functions, Springer, Berlin, 1987. (1987) 
  19. 10.1007/BF00773667, Ann. Inst. Statist. Math. 45 (1993), 35–54. (1993) MR1220289DOI10.1007/BF00773667
  20. Global statistical information, likelihood ratio tests and maximum likelihood estimators, Kybernetika (submitted) (1997). (1997) 
  21. 10.1080/02331889508802499, Statistics 26 (1995), 307–328. (1995) MR1365681DOI10.1080/02331889508802499

NotesEmbed ?

top

You must be logged in to post comments.