Computational Epidemiology: Bayesian Disease Surveillance

1
Computational EpidemiologyBayesian Disease
Surveillance

• Kaja Abbas, Armin R. Mikler,
• Amir Ramezani, Sheena Menezes
• University of North Texas

2
Presentation Outline

• Computational epidemiology
• Brief overview of past epidemics
• Mathematical epidemiological models
• Bayesian learning
• Bayesian disease outbreak model
• Results and analysis
• Conclusion and future work

3
Computational Epidemiology
4
Epidemic History

• 14th century Europe 25 million Plague
• 1521 Aztecs 3.5 million Small pox
• 1918 Worldwide at-least 20 million Influenza
• 2003 SARS rapid global spread

5
Influenza Infection Timeline
6
Mathematical Epidemiology

• Susceptibles Infectives Removals (SIR) model

SIR State Diagram
7
SIR Epidemic Curve
Disease Prevalence
8
Limitations of SIR model

• Explicit representation of outbreak data
• Lack of demographic analysis

Bayesian Epidemic Model

• Learning
• Implicit representation of population
  demographics

9
Bayesian Learning

• Probabilistic Reasoning
• Reasoning under uncertainty

10
Disease Outbreak Bayesian Network
11
Bayesian Network for Demographic Analysis
12
Geographic Region I Probability distributions
13
Geographic Region I Bayesian Network
14
Geographic Region I Analysis
15
Geographic Region II Probability distributions
16
Geographic Region II Bayesian Network
17
Geographic Region II Analysis
18
Regional Comparison of Inferences
Region I
Region II
19
Summary

• SIR Model
• Lack of demographic analysis
• Bayesian Epidemic Model
• Demographic analysis
• Spatial portability of inferences

20
Current Work
21
Discussion, Questions Comments

• Computational Epidemiology Bayesian
  Disease Surveillance
• Kaja Abbas, Armin R. Mikler,
• Amir Ramezani, Sheena Menezes
• Computational Epidemiology Research Laboratory
  (cerl.unt.edu)
• Department of Computer Science and Engineering
• University of North Texas
• Email kaja, mikler_at_cs.unt.edu, ar0116,
  srm0034_at_unt.edu