Functional Approximations to Likelihoods/Posterior Densities: A Neural Network Approach to Efficient Sampling
The performance of Monte Carlo integration methods like importance-sampling or Markov-Chain Monte-Carlo procedures depends greatly on the choice of the importance- or candidate-density. Such a density must typically be "close" to the target density to yield numerically accurate results with efficient sampling. Neural networks are natural importance- or candidate-densities since they have a universal approximation property and are easy to sample from. That is, conditional upon the specified neural network, sampling can be done either directly or using a Gibbs sampling technique, possibly with auxiliary variables. We propose such a class of methods, a key step for which is the construction of a neural network that approximates the target density accurately. The methods are tested on a set of illustrative models that includes a mixture of normal distributions, a Bayesian instrumental-variable regression problem with weak instruments and near-identification, and a two-regime growth model for US recessions and expansions. These examples involve experiments with non-standard, non-elliptical posterior distributions. The results indicate the feasibility of the neural network approach
References listed on IDEAS
Please report citation or reference errors to , or , if you are the registered author of the cited work, log in to your RePEc Author Service profile, click on "citations" and make appropriate adjustments.:
- Neil Shephard, 2005. "Stochastic Volatility," Economics Papers 2005-W17, Economics Group, Nuffield College, University of Oxford.
- Paap, Richard & van Dijk, Herman K, 2003.
"Bayes Estimates of Markov Trends in Possibly Cointegrated Series: An Application to U.S. Consumption and Income,"
Journal of Business & Economic Statistics,
American Statistical Association, vol. 21(4), pages 547-563, October.
- Paap, R. & van Dijk, H.K., 2002. "Bayes estimates of Markov trends in possibly cointegrated series: an application to US consumption and income," Econometric Institute Research Papers EI 2002-42, Erasmus University Rotterdam, Erasmus School of Economics (ESE), Econometric Institute.
- Richard Paap & Herman K. van Dijk, 1999. "Bayes Estimates of Markov Trends in possibly Cointegrated Series: An Application to US Consumption and Income," Tinbergen Institute Discussion Papers 99-024/4, Tinbergen Institute.
- Kleibergen, Frank & van Dijk, Herman K., 1998.
"Bayesian Simultaneous Equations Analysis Using Reduced Rank Structures,"
Cambridge University Press, vol. 14(06), pages 701-743, December.
- Frank Kleibergen & Herman K. van Dijk, 1998. "Bayesian Simultaneous Equations Analysis using Reduced Rank Structures," Tinbergen Institute Discussion Papers 98-025/4, Tinbergen Institute.
- Kleibergen, F.R. & van Dijk, H.K., 1997. "Bayesian Simultaneous Equations Analysis using Reduced Rank Structures," Econometric Institute Research Papers EI 9714/A, Erasmus University Rotterdam, Erasmus School of Economics (ESE), Econometric Institute.
- Hamilton, James D, 1989. "A New Approach to the Economic Analysis of Nonstationary Time Series and the Business Cycle," Econometrica, Econometric Society, vol. 57(2), pages 357-384, March.
- Pesaran, M. Hashem & Smith, Ron, 1995.
"Estimating long-run relationships from dynamic heterogeneous panels,"
Journal of Econometrics,
Elsevier, vol. 68(1), pages 79-113, July.
- Pesaran, M.H. & Smith, R., 1992. "Estimating Long-Run Relationships From Dynamic Heterogeneous Panels," Cambridge Working Papers in Economics 9215, Faculty of Economics, University of Cambridge.
- Schotman, Peter & van Dijk, Herman K., 1991. "A Bayesian analysis of the unit root in real exchange rates," Journal of Econometrics, Elsevier, vol. 49(1-2), pages 195-238.
- Bauwens, L. & Bos, C.S. & van Dijk, H.K. & van Oest, R.D., 2002. "Adaptive polar sampling, a class of flexibel and robust Monte Carlo integration methods," Econometric Institute Research Papers EI 2002-27, Erasmus University Rotterdam, Erasmus School of Economics (ESE), Econometric Institute.
- Bauwens, Luc & Lubrano, Michel & Richard, Jean-Francois, 2000. "Bayesian Inference in Dynamic Econometric Models," OUP Catalogue, Oxford University Press, number 9780198773139.
- Kloek, Tuen & van Dijk, Herman K, 1978. "Bayesian Estimates of Equation System Parameters: An Application of Integration by Monte Carlo," Econometrica, Econometric Society, vol. 46(1), pages 1-19, January.
- Rombouts, Jeroen V. K. & Bauwens, Luc, 2004.
2004,33, Humboldt-Universität Berlin, Center for Applied Statistics and Economics (CASE).
- Chib, Siddhartha & Greenberg, Edward, 1996.
"Markov Chain Monte Carlo Simulation Methods in Econometrics,"
Cambridge University Press, vol. 12(03), pages 409-431, August.
- Siddhartha Chib & Edward Greenberg, 1994. "Markov Chain Monte Carlo Simulation Methods in Econometrics," Econometrics 9408001, EconWPA, revised 24 Oct 1994.
- Geweke, John, 1989. "Bayesian Inference in Econometric Models Using Monte Carlo Integration," Econometrica, Econometric Society, vol. 57(6), pages 1317-1339, November.
- Kleibergen, Frank & van Dijk, Herman K., 1994. "On the Shape of the Likelihood/Posterior in Cointegration Models," Econometric Theory, Cambridge University Press, vol. 10(3-4), pages 514-551, August.
- repec:cup:etheor:v:12:y:1996:i:3:p:409-31 is not listed on IDEAS
- van Dijk, H. K. & Kloek, T., 1980. "Further experience in Bayesian analysis using Monte Carlo integration," Journal of Econometrics, Elsevier, vol. 14(3), pages 307-328, December.
- Luc Bauwens & Charles S. Bos & Herman K. van Dijk & Rutger D. van Oest, 2002. "Adaptive Polar Sampling," Computing in Economics and Finance 2002 307, Society for Computational Economics.
When requesting a correction, please mention this item's handle: RePEc:sce:scecf4:74. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Christopher F. Baum)
If references are entirely missing, you can add them using this form.