Modeling the Temporal Trend of the Daily Severity of an Outbreak Using Bayesian Networks

Abstract

A disease outbreak is an epidemic limited to localized increase, e.g., in a village, town, or institution. An epidemic curve is a graphical depiction of the number of outbreak cases by date of onset of illness. If we could estimate the epidemic curve early in an outbreak, this estimate could guide the investigation of other outbreak characteristics. Furthermore, a good estimate of the epidemic curves tells us how soon the outbreak will reach a given level of severity if it goes uncontrolled. Previously, methods for doing real-time estimation and prediction of the severity of an outbreak were very limited. As far as predicting future cases, ordinarily epidemiologists simply made an educated guess as to how many people might become affected. We develop a Bayesian network model for real-time estimation of an epidemic curve, and we show results of experiments testing its accuracy.