×

Testing order constraints: qualitative differences between Bayes factors and normalized maximum likelihood. (English) Zbl 1396.62042

Summary: We compared Bayes factors to normalized maximum likelihood for the simple case of selecting between an order-constrained versus a full binomial model. This comparison revealed two qualitative differences in testing order constraints regarding data dependence and model preference.

MSC:

62F15 Bayesian inference
PDFBibTeX XMLCite
Full Text: DOI arXiv

References:

[2] Grünwald, P., The Minimum Description Length Principle (2007), MIT Press: MIT Press Cambridge, MA
[3] Kakade, S. M.; Seeger, M. W.; Foster, D. P., Worst-case bounds for Gaussian process models, (Weiss, Y.; Schölkopf, B.; Platt, J., Advances in Neural Information Processing Systems, Vol. 18 (2006), MIT Press: MIT Press Cambridge, MA), 619-626
[4] Kass, R. E.; Raftery, A. E., Bayes factors, J. Amer. Statist. Assoc., 90, 773-795 (1995) · Zbl 0846.62028
[5] Klugkist, I.; Hoijtink, H., The Bayes factor for inequality and about equality constrained models, Comput. Statist. Data Anal., 51, 6367-6379 (2007) · Zbl 1445.62049
[6] Lewis, S. M.; Raftery, A. E., Estimating Bayes factors via posterior simulation with the Laplace-Metropolis estimator, J. Amer. Statist. Assoc., 92, 648-655 (1997) · Zbl 0889.62018
[7] Rissanen, J., Modeling by shortest data description, Automatica, 14, 465-471 (1978) · Zbl 0418.93079
[8] Rissanen, J., Fisher information and stochastic complexity, IEEE Trans. Inf. Theory, 42, 40-47 (1996) · Zbl 0856.94006
[9] Rissanen, J., Strong optimality of the normalized ML models as universal codes and information in data, IEEE Trans. Inform. Theory, 47, 1712-1717 (2001) · Zbl 0999.94016
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.