This paper introduces a class of graphical independence models that is closed under marginalization and conditioning but that contains all DAG independence models. This class of graphs, called maximal ancestral graphs, has two attractive features: there is at most one edge between each pair of vertices; every missing edge corresponds to an independence relation. These features lead to a simple parameterization of the corresponding set of distributions in the Gaussian case.

Introduction The introduction of Bayesian networks (Pearl 1986b) and associated local computation algorithms (Lauritzen and Spiegelhalter 1988, Shenoy and Shafer 1990, Jensen, Lauritzen and Olesen 1990) has initiated a renewed interest for understanding causal concepts in connection with modelling complex stochastic systems. It has become clear that graphical models, in particular those based upon directed acyclic graphs, have natural causal interpretations and thus form a base for a language in which causal concepts can be discussed and analysed in precise terms. As a consequence there has been an explosion of writings, not primarily within mainstream statistical literature, concerned with the exploitation of this language to clarify and extend causal concepts. Among these we mention in particular books by Spirtes, Glymour and Scheines (1993), Shafer (1996), and Pearl (2000) as well as the collection of papers in Glymour and Cooper (1999). Very briefly, but fundamentally,

The class of chain graphs (CGs) involving both undirected graphs (= Markov networks) and directed acyclic graphs (= Bayesian networks) was introduced in middle eighties for description of probabilistic conditional independence structures. Every class of Markov equivalent CGs (that is CGs describing the same conditional independence structure) has a natural representative, which is called the largest CG. The paper presents socalled recovery algorithm, which on basis of the conditional independence structure given by a CG (in form of so-called dependency model) finds the largest CG, representing the corresponding class of Markov equivalent CGs. As a byproduct a graphical characterization of graphs, which are the largest CGs (for a class of Markov equivalent CGs) is obtained, and a simple algorithm changing every CG into the largest CG of the corresponding equivalence class is given. 1 INTRODUCTION Classic graphical approaches to description of probabilistic conditional independence stru...

This paper introduces

The dependence of coincidence of the global, local and pairwise Markov properties on the underlying undirected graph is examined. The pairs of these properties are found to be equivalent for graphs with some small exluded subgraphs. Probabilistic representations of the corresponding conditional independence structures are discussed. MARKOV RANDOM FIELDS, MARKOV PROPERTIES, CONDITIONAL INDEPENDENCE 1. Introduction. Conditional independence restrictions of the Markov type together with factorizations of probability distributions of Markov fields have been investigated over two decades. It is well known that if the distribution factorizes with respect to an undirected graph then it has the corresponding global Markov property which implies the local Markov property yet stronger than the pairwise one [see Lauritzen (1989),Whittaker (1990)]. Alternatively, under a positivity assumption on the density of the distribution a famous result, ascribed to Hammersley and Clifford, asserts the equi...

this paper we discuss maximum likelihood estimation when some observations are missing in mixed graphical interaction models assuming a conditional Gaussian distribution as introduced by Lauritzen & Wermuth (1989). For the saturated case ML estimation with missing values via the EM algorithm has been proposed by Little & Schluchter (1985). We expand their results to the special restrictions in graphical models and indicate a more efficient way to compute the E--step. The main purpose of the paper is to show that for certain missing patterns the computational effort can considerably be reduced. Some key words: EM algorithm; Graphical interaction models; Maximum likelihood estimation; Missing pattern; Missing values. 1. Introduction

Orthogeodesic models admit marginal local cuts and therefore separate inference on subparameters is asymptotically justified. Doubly-flat orthogeodesic models admit local cuts marginally and conditionally. Two important empirical models for panel data are used to illustrate this property and demonstrate its usefulness. The relation to local ancillarity and local sufficiency is explored. An alternative characterization of local cuts in terms of curvature is given and shown to be intrinsic. Applications to semiparametric estimation are considered

The theoretical background for a program for establishing expert systems on the basis of observations and expert knowledge is presented. Block recursive models form the basis of the statistical modelling. These models, together with various model selection methods for automatic model selection, are presented. Additionally, the connection between a block recursive model and expert systems based on causal probabilistic networks is treated. A medical example concerning diagnosis of coronary artery disease forms the basis for an evaluation of the expert systems established. Keywords: causal probabilistic networks, graphical association models, machine learning, model selection, selection criteria, selection strategies. 1 Introduction BIFROST is a program for semi-automatic knowledge acquisition and is a continuation developments made in (Greve, Hjsgaard, Skjth and Thiesson 1990). The objective is to obtain preliminary causal models for use in the HUGIN expert system shell (Ander...

this paper, a class of multivariate distributions introduced by Koehler and Symanowski (1995) is discussed with regard to whether it can be reasonably applied in the framework of graphical modeling. Therefore, the focus lies on properties like marginal and conditional independence, marginalization and the exibility as far as the modeling of a dependence structure is concerned. 1 Introduction and Notations

The combination of graphical models and reference analysis represents a powerful tool for Bayesian inference in highly multivariate settings. It is typically difficult to derive reference priors in complex problems. In this paper we present a suitable mixed parameterisation for a discrete decomposable graphical model and derive the corresponding reference prior.