Abstract
In this paper, we propose Bayesian estimation of unknown parameters for a multinomial logit model. Since the multinomial logit model does not have conjugate prior distribution, we apply a Metropolis-Hastings (MH) algorithm. The MH algorithm requires prior and proposal distributions, which are frequently chosen according to the researcher’s belief or subjectivity. Unless we give appropriate distributions, the result is an inefficient or unbalanced estimation. Hence, we propose a combined empirical prior distribution that assumes the independence of the prior distribution for each of the parameter groups with an estimation procedure to obtain the appropriate estimation. We confirm the validity of the proposal method using multiple convergence diagnostics.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Bishop, C. M.: Pattern Recognition and Machine Learning. Springer Texts in Statistics. New York, 2006.
Jeffreys, H.: Theory of Probability. Oxford University Press, 1961
Zellner, A.: An Introduction to Bayesian Inference in Econometrics. John Wiley and Sons, 1971
Kass R., Wasserman L.: The selection of prior distributions by formal rules. Journal of the American Statistical Association. 91, 1343–1370 (1996)
Tierney L.: Markov chains for exploring posterior distributions. The Annals of Statistics, 22, 1701–1762 (1994)
Geman S., Geman D.: Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images. IEEE Transactions on Pattern Analysis and Machine Intelligence. 6, 721–741 (1984)
Metropolis N., Rosenbluth A. W., Rosenbluth M. N., Teller A. H., Teller E.: Equation of state calculations by fast computing machines. Journal of Chemical Physics. 21, 1087–1092 (1953)
Hastings W. K.: Monte Carlo sampling methods using Markov chains and their applications. Biometrika. 57, 97–109 (1970)
Holmes C. C., Held L.: Bayesian auxiliary variable models for binary and multinomial regression. Bayesian Analysis. 1(1), 145–168 (2006)
Albert J. H., Chib S.: Bayesian analysis of binary and polychotomous response data. Journal of the American Statistical Association. 88, 669–679 (1993)
Rossi, P., Allenby, G. M., McCulloch, R.: Bayesian Statistics and Marketing. Wiley, 2005
Gelman A., Roberts G. O.: Gilks W.R.: Efficient Metropolis jumping rules. Bayesian Statistics. 5, 599–607 (1996)
Roberts G. O., Gelman A., Gilks W. R.: Weak convergence and optimal scaling of random walk Metropolis algorithms. The Annals of Applied Probability. 7, 110–120 (1997)
Kass R., Carlin B., Gelman A., Neal R.: Markov chain Monte Carlo in practice: A roundtable discussion. The American Statistician. 52, 93–100 (1998)
Hoff, P. D.: A First Course in Bayesian Statistical Methods. Springer Texts in Statistics. New York, 2009
Geweke J.: Evaluating the accuracy of sampling–based approaches to the calculation of posterior moments. Bayesian Statistics. 4, 169–193 (1992)
Gelman, A.: Inference and monitoring convergence. In W. R. Gilks, S. Richardson, and D. J. Spiegelhalter eds. Markov Chain Monte Carlo in Practice, pages 131–143. Chapman and Hall/CRC, London, 1996
Japan Meteorological Agency. http://www.jma.go.jp/jma/indexe.html.
Matsumoto, N., Kurosawa, T.: Bayesian estimations with isotropic and anisotropic matrices for a multinomial logit model. In Procedia Computer Science, volume 35, pages 1474–1481. Elsevier, 2014. 18th International Conference on Knowledge-Based and Intelligent Information and Engineering Systems.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Matsumoto, N., Kurosawa, T. Bayesian Estimation with Combined Empirical Prior Distribution for a Multinomial Logit Model. Rev Socionetwork Strat 9, 59–74 (2015). https://doi.org/10.1007/s12626-015-0056-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s12626-015-0056-1