Abstract
Coherent reasoning under uncertainty can be represented in a very general manner by coherent sets of desirable gambles. In this framework, and for a given finite category set, coherent predictive inference under exchangeability can be represented using Bernstein coherent cones of multivariate polynomials on the simplex generated by this category set. This is a powerful generalisation of de Finetti’s representation theorem allowing for both imprecision and indecision. We define an inference system as a map that associates a Bernstein coherent cone of polynomials with every finite category set. Many inference principles encountered in the literature can then be interpreted, and represented mathematically, as restrictions on such maps. We discuss two important inference principles: representation insensitivity—a strengthened version of Walley’s representation invariance—and specificity. We show that there is a infinity of inference systems that satisfy these two principles, amongst which we discuss in particular the inference systems corresponding to (a modified version of) Walley and Bernard’s imprecise Dirichlet multinomial models (IDMMs) and the Haldane inference system.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
… unless the observed sequence has probability zero.
- 2.
To avoid confusion, we make a (perhaps non-standard) distinction between the multinomial expectation, which is associated with sequences of observations, and the count multinomial expectation, associated with their count vectors.
- 3.
The degree may be smaller than n because the sum of all Bernstein basis polynomials of fixed degree is one. Strictly speaking, these polynomials p are restrictions to \(\Sigma_{A}\) of multivariate polynomials q on \(\mathbb{R}^A\), called representations of p. For any p, there are multiple representations, with possibly different degrees. The smallest such degree is then called the degree deg\((p)\) of p.
- 4.
Strictly speaking, Eq. 2.7 only defines the count multinomial expectation operator \(\text{CoMn}^{n}_{A}\) for \(n>0\), but it is clear that the definition extends trivially to the case n = 0.
- 5.
Actually, a suitably adapted version of coherence, where the gambles are restricted to the polynomials on \(\Sigma_{A}\).
- 6.
It is an immediate consequence of the F. Riesz extension theorem that each such linear prevision is the restriction to polynomials of the expectation operator of some unique σ-additive probability measure on the Borel sets of \(\Sigma_{A}\); see for instance [6].
References
Augustin, T., Coolen, F.P.A., de Cooman, G., Troffaes, M.C.M. (eds.): Introduction to Imprecise Probabilities. Wiley (2014)
Bernard, J.M.: Bayesian analysis of tree-structured categorized data. Revue Internationale de Systémique 11, 11–29 (1997)
Bernard, J.M.: An introduction to the imprecise Dirichlet model for multinomial data. Int. J. Approx. Reason 39, 123–150 (2005)
Cifarelli, D.M., Regazzini, E.: De Finetti’s contributions to probability and statistics. Stat. Sci. 11, 253–282 (1996)
Couso, I., Moral, S.: Sets of desirable gambles: conditioning, representation, and precise probabilities. Int. J. Approx. Reason 52(7), 1034–1055 (2011)
de Cooman, G., Miranda, E.: The F. Riesz representation theorem and finite additivity. In: D. Dubois, M.A. Lubiano, H. Prade, M.A. Gil, P. Grzegorzewski, O. Hryniewicz (eds.) Soft Methods for HandlingVariability and Imprecision (Proceedings of SMPS 2008), pp. 243–252. Springer (2008)
de Cooman, G., Miranda, E.: Irrelevant and independent natural extension for sets of desirable gambles. J. Artif. Intell. Res. 45, 601–640 (2012). http://www.jair.org/vol/vol45.html
de Cooman, G., Quaeghebeur, E.: Exchangeability and sets of desirable gambles. Int. J. Approx. Reason 53(3), 363–395 (2012). (Special issue in honour of Henry E. Kyburg, Jr.).
de Cooman, G., Miranda, E., Quaeghebeur, E.: Representation insensitivity in immediate prediction under exchangeability. Int. J. Approx. Reason 50(2), 204–216 (2009). doi:10.1016/j.ijar.2008.03.010
de Cooman, G., Quaeghebeur, E., Miranda, E.: Exchangeable lower previsions. Bernoulli 15(3), 721–735 (2009). doi:10.3150/09-BEJ182. http://hdl.handle.net/1854/LU-498518
de Finetti, B.: La prévision: ses lois logiques, ses sources subjectives. Annales de l’Institut Henri Poincaré 7, 1–68 (1937). (English translation in [18])
de Finetti, B.: Teoria delle Probabilità. Einaudi, Turin (1970)
de Finetti, B.: Theory of Probability: A Critical Introductory Treatment. Wiley, Chichester (1974–1975). (English translation of [12], two volumes)
Haldane, J.B.S.: On a method of estimating frequencies. Biometrika 33, 222–225 (1945)
Jeffreys, H.: Theory of Probability. Oxford Classics series. Oxford University Press (1998). (Reprint of the third edition (1961), with corrections)
Johnson, N.L., Kotz, S., Balakrishnan, N.: Discrete Multivariate Distributions.Wiley Series in Probability and Statistics.Wiley, NewYork (1997)
Jaynes, E.T.: Probability Theory: The Logic of Science. Cambridge University Press (2003)
Kyburg Jr., H.E., Smokler, H.E. (eds.): Studies in Subjective Probability. Wiley, New York (1964). (Second edition (with new material) 1980 )
Lad, F.: Operational Subjective Statistical Methods: A Mathematical, Philosophical and Historical Introduction. Wiley (1996)
Mangili, F., Benavoli, A.: New prior near-ignorance models on the simplex. In: F. Cozman, T. Denoeux, S. Destercke, T. Seidenfeld (eds.) ISIPTA '13 – Proceedings of the Eighth International Symposium on Imprecise Probability: Theories and Applications, pp. 213–222. SIPTA (2013)
Moral, S.: Epistemic irrelevance on sets of desirable gambles. Ann. Math. Artif. Intell. 45, 197–214 (2005). doi:10.1007/s10472-005-9011-0
Quaeghebeur, E.: Introduction to Imprecise Probabilities, (Chapter: Desirability). Wiley (2014)
Quaeghebeur, E., de Cooman, G., Hermans, F.: Accept & reject statement-based uncertainty models. Int. J. Approx. Reason. (2013). (Submitted for publication)
Rouanet, H., Lecoutre, B.: Specific inference in ANOVA: From significance tests to Bayesian procedures. Br. J. Math. Stat. Psychol. 36(2), 252–268 (1983). doi:10.1111/j.2044-8317.1983.tb01131.x
Walley, P.: Statistical Reasoning with Imprecise Probabilities. Chapman and Hall, London (1991)
Walley, P.: Inferences from multinomial data: learning about a bag of marbles. J. R. Stat. Soc., Series B 58, 3–57 (1996). (With discussion)
Walley, P.: A bounded derivative model for prior ignorance about a real-valued parameter. Scand. J. Stat. 24(4), 463–483 (1997). doi:10.1111/1467-9469.00075
Walley, P.: Towards a unified theory of imprecise probability. Int. J. Approx. Reason 24, 125–148 (2000)
Walley, P., Bernard, J.M.: Imprecise probabilistic prediction for categorical data. Tech. Rep. CAF-9901, Laboratoire Cognition et Activitées Finalisées, Université de Paris 8 (1999)
Williams, P.M.: Indeterminate probabilities. In: M. Przelecki, K. Szaniawski, R. Wojcicki (eds.) Formal Methods in the Methodology of Empirical Sciences, pp. 229–246. Reidel, Dordrecht (1976). (Proceedings of a 1974 conference held inWarsaw)
Acknowledgements
Gert de Cooman’s research was partially funded through project number 3G012512 of the Research Foundation Flanders (FWO). Jasper De Bock is a PhD Fellow of the Research Foundation Flanders and wishes to acknowledge its financial support. Marcio Diniz was supported by FAPESP (São Paulo Research Foundation), under the project 2012/14764-0 and wishes to thank the SYSTeMS Research Group at Ghent University for its hospitality and support during his sabbatical visit there.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2015 Springer International Publishing Switzerland
About this paper
Cite this paper
de Cooman, G., De Bock, J., Diniz, M. (2015). Predictive Inference Under Exchangeability, and the Imprecise Dirichlet Multinomial Model. In: Polpo, A., Louzada, F., Rifo, L., Stern, J., Lauretto, M. (eds) Interdisciplinary Bayesian Statistics. Springer Proceedings in Mathematics & Statistics, vol 118. Springer, Cham. https://doi.org/10.1007/978-3-319-12454-4_2
Download citation
DOI: https://doi.org/10.1007/978-3-319-12454-4_2
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-12453-7
Online ISBN: 978-3-319-12454-4
eBook Packages: Mathematics and StatisticsMathematics and Statistics (R0)