Abstract
In this paper, we first propose a noniterative sampling method to obtain an i.i.d. sample approximately from posteriors by combining the inverse Bayes formula, sampling/importance resampling and posterior mode estimates. We then propose a new exact algorithm to compute posteriors by improving the PMDA-Exact using the sampling-wise IBF. If the posterior mode is available from the EM algorithm, then these two algorithms compute posteriors well and eliminate the convergence problem of Markov Chain Monte Carlo methods. We show good performances of our methods by some examples.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Chen MH, Shao QM and Ibrahim JG (2000). Monte Carlo methods in Bayesian computation. Springer, New York
Dempster AP, Laird NM and Rubin DB (1977). Maximum likelihood from incomplete data via the EM algorithm (with discussion). J R Stat Soc Ser B39: 1–38
Gelfand AE (2002) Gibbs sampling. In: Rafery AE, Tanner MA, Wells MT (eds) Statistics in the twenty-first century. Chapman and Hall/CRC, Boca Raton, pp 341–349
Jones GL and Hobert JP (2001). Honest exploration of intractable probability distribution via Markov chain Monte Carlo. Stat Sci 16: 312–334
Kuo L and Yang TY (1996). Bayesian computation for nonhomogeneous Poisson process in software reliability. J Am Stat Assoc 91: 763–773
Laird NM and Ware JH (1982). Random effects models for longitudinal data. Biometrics 38: 963–974
Liu CH and Rubin DB (1994). The ECME algorithm: A simple extension of EM and ECM with faster monotone convergence. Biometrika 81: 633–648
Ng KW (1995) Explicit formulas for unconditional pdf. Research Report, No. 82 (revised), Department of Statistics, University of Hong Kong
Ng KW (1997) Inversion of Bayes formula: explicit formulas for unconditional pdf. In: Johnson NL, Balakrishnan N (eds). Advances in the theory and practice in statistics. Wiley, New York, pp 571–584
Robert CP and Casella G (1999). Monte Carlo statistical methods. Springer, New York
Rubin DB (1988) Using the SIR algorithm to simulate posterior distributions (with discussion). In: Berardo JM, DeGroot MH, Lindley DV, Smith AFM (eds). Bayes statistics vol 3. Oxford University Press, Oxford, pp 395–402
Tan M, Tian GL and Ng KW (2003). A noniterative sampling method for computing posteriors in the structure of EM-type algorithms. Stat Sin 13: 625–639
Tanner MA (1992). Lecture notes in statistics. Springer, New York
Tanner MA and Wong WH (1987). The calculation of posterior distributions by data augmentation (with discussion). J Am Stat Assoc 82: 528–540
Tian GL, Tan M (2003) Exact statistical solutions using the inverse Bayes formulae. Stat Probab Lett 62:305–315
Tierney L and Kadane JB (1986). Accurate approximations for posterior moments and marginal densities. J Am Stat Assoc 81: 82–86
Tierney L, Kass RE and Kadane JB (1989). Fully exponential Laplace approximations expectations and variances of nonpositive functions. J Am Stat Assoc 84: 710–716
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Yang, J., Zou, G. & Zhao, Y. Two noniterative algorithms for computing posteriors. Comput Stat 23, 443–453 (2008). https://doi.org/10.1007/s00180-007-0085-5
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-007-0085-5