Bayesian networks for cardiovascular monitoring

Abstract

In the Intensive Care Unit, physicians have access to many types of information when treating patients. Physicians attempt to consider as much of the relevant information as possible, but the astronomically large amounts of data collected make it impossible to consider all available information within a reasonable amount of time. In this thesis, I explore Bayesian Networks as a way to integrate patient data into a probabilistic model. I present a small Bayesian Network model of the cardiovascular system and analyze the network's ability to estimate unknown patient parameters using available patient information. I test the network's estimation capabilities using both simulated and real patient data, and I discuss ways to exploit the network's ability to adapt to patient data and learn relationships between patient variables.