Semigraphoids and structures of probabilistic conditional independence

The concept of conditional independence (CI) has an important role in probabilistic reasoning, that is a branch of artificial intelligence where knowledge is modeled by means of a multidimensional finite-valued probability distribution. The structures of probabilistic CI are described by means of semigraphoids, that is lists of CI-statements closed under four concrete inference rules, which have at most two antecedents. It is known that every CI-model is a semigraphoid, but the converse is not true. In this paper, the semigraphoid closure of every couple of CI-statements is proved to be a CI-model. The substantial step to it is to show that every probabilistically sound inference rule for axiomatic characterization of CI properties (= axiom), having at most two antecedents, is a consequence of the semigraphoid inference rules. Moreover, all potential dominant triplets of the mentioned semigraphoid closure are found.

[1]  Dan Geiger,et al.  Axioms and Algorithms for Inferences Involving Probabilistic Independence , 1991, Inf. Comput..

[2]  A Perez,et al.  Constructing an intentional expert system , 1985 .

[3]  Wolfgang Spohn,et al.  Stochastic independence, causal independence, and shieldability , 1980, J. Philos. Log..

[4]  Frantisek Matús.  Stochastic Independence, Algebraic Independence and Abstract Connectedness , 1994, Theor. Comput. Sci..

[5]  Yehoshua Sagiv,et al.  Subset Dependencies and a Completeness Result for a Subclass of Embedded Multivalued Dependencies , 1982, JACM.

[6]  Francesco M. Malvestuto.  A unique formal system for binary decompositions of database relations, probability distributions, and graphs , 1992, Inf. Sci..

[7]  Prakash P. Shenoy,et al.  Conditional independence in valuation-based systems , 1994, Int. J. Approx. Reason..

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

[9]  Milan Studený.  Formal Properties of Conditional Independence in Different Calculi of AI , 1993, ECSQARU.

[10]  J. Q. Smith.  Influence Diagrams for Statistical Modelling , 1989 .

[11]  F. Matús.  PROBABILISTIC CONDITIONAL INDEPENDENCE STRUCTURES AND MATROID THEORY: BACKGROUND1 , 1993 .

[12]  Francesco M. Malvestuto,et al.  Comment on "A unique formal system for binary decompositions of database relations, probability distributions, and graphs" , 1992, Inf. Sci..

[13]  A. Dawid.  Conditional Independence in Statistical Theory , 1979 .

[14]  C. V. D. Gaag.  Characterising Normal Forms for Informational IndependenceL , 1996 .

[15]  Alan Bundy,et al.  Symbolic and Quantitative Approaches to Reasoning and Uncertainty , 1993 .

[16]  Judea Pearl,et al.  Probabilistic reasoning in intelligent systems , 1988 .

[17]  N. Wermuth,et al.  Graphical Models for Associations between Variables, some of which are Qualitative and some Quantitative , 1989 .

[18]  Judea Pearl,et al.  GRAPHOIDS: A Graph-based logic for reasoning about relevance relations , 1985 .

[19]  Milan Studeny,et al.  Conditional independence relations have no finite complete characterization , 1992 .

[20]  Uwe Franz.  What is Stochastic Independence , 2002 .

[21]  T. Speed,et al.  Recursive causal models , 1984, Journal of the Australian Mathematical Society. Series A. Pure Mathematics and Statistics.

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

[23]  T. Speed,et al.  Markov Fields and Log-Linear Interaction Models for Contingency Tables , 1980 .

[24]  Milan Studený,et al.  Semigraphoids Are Two-Antecedental Approximations of Stochastic Conditional Independence Models , 1994, UAI.

[25]  J. Pearl,et al.  Logical and Algorithmic Properties of Conditional Independence and Graphical Models , 1993 .