Log In Sign Up

Learning Bayesian Networks with Restricted Causal Interactions

by   Julian R. Neil, et al.

A major problem for the learning of Bayesian networks (BNs) is the exponential number of parameters needed for conditional probability tables. Recent research reduces this complexity by modeling local structure in the probability tables. We examine the use of log-linear local models. While log-linear models in this context are not new (Whittaker, 1990; Buntine, 1991; Neal, 1992; Heckerman and Meek, 1997), for structure learning they are generally subsumed under a naive Bayes model. We describe an alternative interpretation, and use a Minimum Message Length (MML) (Wallace, 1987) metric for structure learning of networks exhibiting causal independence, which we term first-order networks (FONs). We also investigate local model selection on a node-by-node basis.


page 1

page 2

page 5

page 6

page 7


On tensor rank of conditional probability tables in Bayesian networks

A difficult task in modeling with Bayesian networks is the elicitation o...

Learning Bayesian Networks with Local Structure

In this paper we examine a novel addition to the known methods for learn...

Embedded Bayesian Network Classifiers

Low-dimensional probability models for local distribution functions in a...

Assessing Bayesian Nonparametric Log-Linear Models: an application to Disclosure Risk estimation

We present a method for identification of models with good predictive pe...

Categorical Aspects of Parameter Learning

Parameter learning is the technique for obtaining the probabilistic para...

Technical properties of Ranked Nodes Method

This paper presents analytical and experimental results on the ranked no...