Epid 814 - Advanced Modeling Methods

Course Syllabus

• Syllabus

---

Course schedule

To be added after the initial topic discussion on Monday. Topics are subject to change and rearrangement! (We may also not get to all the topics listed)

1. Introduction and topics discussion (Aug 30, 2021)

List of potential topics to cover in the course - please feel free to add! We will discuss on Monday

2. Parameter estimation and maximum likelihood (Sep 1, 2021 & Sep 8, 2021)

• Maximum likelihood with mathematical models - Lecture

• Parameter estimation coding session - Lab

• Model vs. code comparison

Journal Club (Sep 13, 2021)

• Dissecting racial bias in an algorithm used to manage the health of populations
• Reminder to add papers and sign up to lead a journal club here

3. Uncertainty and identifiability

• Identifiability and parameter uncertainty - Lecture

• Useful post on Wald/FIM vs. Likelihood-based confidence intervals

• Identifiability and uncertainty Lab

4. Intro to MCMC & Bayesian approaches

• Intro to Bayesian Approaches to Parameter Estimation, MCMC - Lecture

• Lab - Metropolis Sampler

• BayesianTools Lab

  • Example code from class

Journal Club (Oct 4, 2021)

• This week we’re discussing two recent papers by Jon Zelner’s group!
• There are no equal opportunity infectors: Epidemiological modelers must rethink our approach to inequality in infection risk (Kelly discussion lead)
• Accounting for Uncertainty During a Pandemic
• Reminder to add papers and sign up to lead a journal club here

5. Data-scraping and data-wrangling

(moved this up since it might be useful for making data sets we can play with for the ML and networks sections below)

• Using R to pull data using API’s - Twitter, Facebook, and Google Maps

• Collecting and working with twitter data

• Twitter data collection lab

  • Rtweet code
  • Webpage scraping code

• Additional useful tutorials/code

  • Tutorial for reading text from pdfs into R using the pdftools package
  • Followup tutorial on parsing the data with regular expressions
  • Two tutorials on extracting table data from pdfs using the tabulizer package: tutorial 1, tutorial 2
  • CronR package for scheduling scripts (we can show some examples of this as well in class if people are interested)

• An interesting article about the ethics of using public data and our perception of privacy

Journal Club (Oct 13, 2021)

• A Bayesian Workflow for Parameter Estimation (Hannah discussion lead)

• Reminder to add papers and sign up to lead a journal club here

6. Sampling-based approaches to global sensitivity analysis

• Latin hypercube and Sobol sampling, local and global sensitivity methods

• Lecture Slides

• Example code

• Interesting paper on Sobol vs. LHS: To Sobol or not to Sobol? The effects of sampling schemes in systems biology applications

• Journal club paper: Marino et al. “A methodology for performing global uncertainty and sensitivity analysis in systems biology.” Journal of theoretical biology 254.1 (2008): 178-196.

• Potentially for later: active subspace methods, surrogate models

Journal Club (Oct 25, 2021 and Oct 27, 2021)

• Journal club paper: Marino et al. “A methodology for performing global uncertainty and sensitivity analysis in systems biology.” Journal of theoretical biology 254.1 (2008): 178-196.

• Oct 27: Discuss the ABMs portion of the paper, and try out some of the methods together

• This R package includes the eFAST method described in the journal club paper, plus a ton more: https://cran.r-project.org/web/packages/sensitivity/index.html

• And this package computes Sobol/variance-based sensitivity analyses:https://cran.r-project.org/web/packages/sensobol/vignettes/sensobol.pdf

• Reminder to add papers and sign up to lead a journal club here

7. Basic intro to some useful algorithms/machine learning methods - Clustering

• Clustering Methods Lecture
  • Density-sensitive version of DBSCAN: OPTICS
  • Another overview of DBSCAN and some of its variants
• Clustering Lab
  • Other useful tutorials:
    • Cluster analysis in R
    • Clustering in R Overview
• Brainstorming twitter data collection ideas

8. Reproduction number - next generation matrix approaches

• The next generation matrix and thinking about the linear algebra of the next generation matrix

• Maybe: type and target reproduction numbers, stochastic \(\mathcal{R}_0\), \(\mathcal{R}_0\) for networks

• Guest lecture by Andrew Brouwer

  • Slides
  • Exercises

9. Reproduction number - data driven approaches to \(\mathcal{R}_0\) and \(\mathcal{R}_t\)

• Quick recap: what are \(\mathcal{R}_0\) and \(\mathcal{R}_t\)?

• Serial interval approaches to calculating \(\mathcal{R}_t\)

• Possible journal club with Royal Society review paper

• Slides

• Rt examples with Michigan COVID data

• Rt examples with simulated data

10. Parallelization in R and Python

• Slides

• Code for R

• Code for Python

11. Basic intro to some useful algorithms/machine learning methods - Classification

• Classification
• [Lab: Naive Bayes classification with twitter data]
  • Other useful tutorials on Naive Bayes
    • Naive Bayes in R for movie reviews
    • Naive Bayes with Titanic survivorship
    • Naive Bayes Overview - includes more details on cross validation, confusion matrices, Laplace smoother, etc.

12. Time series analysis intro

13. Visualization with ggplot2

• Intro to ggplot and collection of examples

• ggrepel example

• Nice ggplot tutorial

• Another tutorial

• Mapping with ggplot

• ggdag & dagitty

14. Other topics to cover if time!

• Parallelizing code in R (maybe combine with another topic since we can do an example fairly quickly)

• Shiny apps and interactives

• Bootstrapping

• Methods for stochastic models

• Compartmental models and DAGs

• And other stuff—topological data analysis, Kullback-Liebler divergence

Journal club topics

• Dissecting racial bias in an algorithm used to manage the health of populations

Useful Links

• EECS 545 (Machine Learning) - online course with a good overview of machine learning approaches