A characterization of the Dirichlet distribution through global and local parameter independence

We provide a new characterization of the Dirichlet distribution. Let θ ij , 1 ≤ i < k, 1 ≤ j ≤ n, be positive random variables that sum to unity. Define θ i = Σ j n = 1 θ ij , θ 1 = {θ i } i=1 k=1 , θ j/i = θ ij /Σj θ ij and θ j/i = {θ j/i } j=1 n-1 . We prove that if {θi, θ j/1 ,..., θ j/k } are mutually independent and {θ. j , θ l/1 ,..., θ I/n } are mutually independent (where θ. j and θ I/j are defined analogously), and each parameter set has a strictly positive pdf, then the pdf of θ ij is Dirichlet. This characterization implies that under assumptions made by several previous authors for selecting a Bayesian network structure out of a set of candidate structures, a Dirichlet prior on the parameters is inevitable.

[1]  R. F.,et al.  Mathematical Statistics , 1944, Nature.

[2]  J. Aczél,et al.  Lectures on Functional Equations and Their Applications , 1968 .

[3]  M. Degroot.  Optimal Statistical Decisions , 1970 .

[4]  J. Darroch,et al.  A Characterization of the Dirichlet Distribution , 1971 .

[5]  J. Fabius.  Two Characterizations of the Dirichlet Distribution , 1973 .

[6]  J. Mosimann,et al.  A New Characterization of the Dirichlet Distribution Through Neutrality , 1980 .

[7]  S. Zabell.  W. E. Johnson's "Sufficientness" Postulate , 1982 .

[8]  N. Wermuth,et al.  Graphical and recursive models for contingency tables , 1983 .

[9]  Antal Járai,et al.  On regular solutions of functional equations , 1986 .

[10]  Ross D. Shachter.  Evaluating Influence Diagrams , 1986, Oper. Res..

[11]  Judea Pearl,et al.  Fusion, Propagation, and Structuring in Belief Networks , 1986, Artif. Intell..

[12]  Judea Pearl,et al.  Evidential Reasoning Using Stochastic Simulation of Causal Models , 1987, Artif. Intell..

[13]  David J. Spiegelhalter,et al.  Local computations with probabilities on graphical structures and their application to expert systems , 1990 .

[14]  Ross D. Shachter.  Probabilistic Inference and Influence Diagrams , 1988, Oper. Res..

[15]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems - networks of plausible inference , 1991, Morgan Kaufmann series in representation and reasoning.

[16]  Judea Pearl,et al.  Equivalence and Synthesis of Causal Models , 1990, UAI.

[17]  Dan Geiger,et al.  Identifying independence in bayesian networks , 1990, Networks.

[18]  J. N. R. Jeffers,et al.  Graphical Models in Applied Multivariate Statistics. , 1990 .

[19]  David Heckerman,et al.  Probabilistic similarity networks , 1991, Networks.

[20]  David J. Spiegelhalter,et al.  Sequential updating of conditional probabilities on directed graphical structures , 1990, Networks.

[21]  Wray L. Buntine.  Theory Refinement on Bayesian Networks , 1991, UAI.

[22]  B. Ramachandran,et al.  Functional equations in probability theory , 1991 .

[23]  Judea Pearl,et al.  Belief Networks Revisited , 1993, Artif. Intell..

[24]  David J. Spiegelhalter,et al.  Bayesian analysis in expert systems , 1993 .

[25]  A. Dawid,et al.  Hyper Markov Laws in the Statistical Analysis of Decomposable Graphical Models , 1993 .

[26]  J. Pearl.  Causal diagrams for empirical research , 1995 .

[27]  Calyampudi Radhakrishna Rao,et al.  Choquet-Deny Type Functional Equations with Applications to Stochastic Models , 1995 .

[28]  David Maxwell Chickering,et al.  A Transformational Characterization of Equivalent Bayesian Network Structures , 1995, UAI.

[29]  David Heckerman,et al.  A Characterization of the Dirichlet Distribution with Application to Learning Bayesian Networks , 1995, UAI.