Abstract:
This program provides a comprehensive four-day exploration of epidemiological modeling and analysis. Participants will learn about compartimental models, R₀/R_t analysis, probability, and statistics. They will also address the limitations of basic models, handle uncertainty through Monte Carlo simulations, and construct posterior distributions. The program goes beyond epidemic prediction, covering incidence estimation, survival analysis, and parameter inference. Participants will also learn regression techniques, Koopman analysis, and Dynamic Mode Decomposition. Practical coding demonstrations accompany the learning process, ensuring participants gain valuable skills in disease dynamics analysis
Schedule:
- Monday 29/05, 13:30-15:30, Aula H
- Tuesday 30/05, 10:30-12:30, Aula G
- Wednesday 31/05, 13:30-15:30, Aula H
- Thursday 01/06, 13:30-15:30, Aula T.03 (Ca' Vignal 3)
Program below:
Day1 Basics of epidemiological modeling:
-SIR models and variants
-Analysis of R_0 / R_t
-Basics of probability and statistics
Day2 Reality of epidemiological modeling and analysis
-Why basic models aren't typically sufficient:
-implicit assumptions of SIR-type models
-Dealing with uncertainty: Monte Carlo simulations
-Case 1: Monte Carlo methods and Latin Hypercube methods when you know/have some idea of the distributions
-Case 2: constructing posterior distributions from Monte Carlo procedures
Day3: Beyond epidemic prediction/progression
-Back-calculation and deconvolution methods for incidence estimation
-Survival analysis techniques
-Maximum likelihood approaches for inferring parameters and distributions
Day4: Modeling with no model
-Linear, nonlinear and logistic regression techniques
-Koopman analysis and DMD for general time-series data
-Coding demonstrations
