Inference with Multinomial Data: Why to Weaken the Prior Strength

@inproceedings{Campos2011InferenceWM,
  title={Inference with Multinomial Data: Why to Weaken the Prior Strength},
  author={Cassio Polpo de Campos and Alessio Benavoli},
  booktitle={IJCAI},
  year={2011}
}
This paper considers inference from multinomial data and addresses the problem of choosing the strength of the Dirichlet prior under a meansquared error criterion. We compare the Maximum Likelihood Estimator (MLE) and the most commonly used Bayesian estimators obtained by assuming a prior Dirichlet distribution with “noninformative” prior parameters, that is, the parameters of the Dirichlet are equal and altogether sum up to the so called strength of the prior. Under this criterion, MLE becomes… CONTINUE READING
4 Citations
10 References
Similar Papers

References

Publications referenced by this paper.
Showing 1-10 of 10 references

pages 22–33

  • A. Benavoli, C. P. de Campos. Inference from multinomial data base ECSQARU
  • Springer,
  • 2009
2 Excerpts

pages 334–342

  • C. Zhai, J. Lafferty. A study of smoothing methods for languag Research, development in information retrieval
  • New York, NY, USA,
  • 2001
2 Excerpts

Bayesian reliability analysis using the Dirichlet prior distribution with emphasis on accelerated life testing run in random order

  • I. F. Somerville, D. L. Dietrich, T. A. Mazzuchi
  • Nonlinear Analysis, 30(7):4415–4423
  • 1997
2 Excerpts

Construction of improved estimators in multiparameter estimation for discrete exponential families

  • M. Ghosh, J. T. Hwang, K. W Tsui
  • Ann. Statist., pages 351–376
  • 1983
1 Excerpt

On admissible estimators for certain fixed sample binomial problems

  • B. M. Johnson
  • Ann. Math. Statist., 42:1579–1587
  • 1971
1 Excerpt

Inadmissibility of the usual estimator for the mean of a multivariate normal distribution

  • C. Stein
  • Proc. Third Berkeley Symp. Math. Statist. Prob…
  • 1956
1 Excerpt

Construction of improved estimators in multiparameter estimation for discrete exponential families On admissible estimators for certain fixed sample binomial problems

  • N. Friedman
  • Ann . Math . Statist .

Similar Papers

Loading similar papers…