Probabilistic Modelling with Bayesian Networks

Francesco Sambo, Fulvia Ferrazzi, Riccardo Bellazzi

Research output: Chapter in Book/Report/Conference proceedingChapter

Abstract

This chapter introduces a probabilistic approach to modelling in physiology and medicine: the quantities of interest are modeled as random variables and the focus is on the probabilistic dependencies between these variables. As primary tool in this modelling framework, we present Bayesian networks (BNs), which map the dependencies between a set of random variables to a directed acyclic graph, both increasing human readability and simplifying the representation of the joint probability distribution of the set of variables. The chapter first describes the theoretical foundations of BNs, including a brief review of probability and graph theory, a formal definition of BNs and details on discrete, continuous, and dynamic BNs. Then, a selection of algorithms for inference, conditional probability learning, and structure learning is presented. Finally, several examples of BN applications in biomedicine are reviewed.

Original languageEnglish
Title of host publicationModeling Methodology for Physiology and Medicine: Second Edition
PublisherElsevier Inc.
Pages257-280
Number of pages24
ISBN (Print)9780124115576
DOIs
Publication statusPublished - Dec 2013

Keywords

  • Bayesian networks
  • Conditional probability estimation
  • Graphical models
  • Probabilistic inference
  • Probabilistic modelling
  • Structure learning

ASJC Scopus subject areas

  • Engineering(all)

Fingerprint Dive into the research topics of 'Probabilistic Modelling with Bayesian Networks'. Together they form a unique fingerprint.

Cite this