Applied Statistical Methods

This is a placeholder page for a survey course on different applied statistical topics. A survey course is not meant to give you a deep understanding into any of the topics but to present the challenges that gave rise to the method and

Course material

Syllabi:

• 3105 Applied Statistical Methods

Pre-requisites

• linear regression
  • the assumptions behind using linear regression for inferring the “true” model (if this sentence doesn’t make sense, that means you don’t have it).
    • how to validate these assumptions
  • interpreting the linear regression coefficients
  • uncertainty of the linear model and the issue of multiple hypothesis testing
  • how to simulate the regression model for data generation
  • How to improve regression
    • creating features (polynomials, interactions, etc)
    • feature selection

Computer setting up

I encourage you to set up Jupyter Notebooks on your computer so you could repeat these in R or Python in the future. For Python users, you’ll need numpy and statsmodel.

Possible topics

• Population and sample collection
  • goals in collecting data
    • measure what you want
    • minimize the uncertainty
  • stratified sampling vs cluster sampling
• Refresher on linear regression
  • Diagnosing regression with real data and improvements
• Bayesian view on linear regression
  • graphical models
• Time series data
  • Kalman filters
  • Contrast with linear regression
• Kriging
  • Estimating unknown surfaces
  • Mining data?
• Causal inference
  • Counter factual and potential outcomes
  • Randomized experiments
  • Propensity scores
  • Synthetic controls
• Survival analysis
  • censor issues
• Causal inference using observational studies
• False Discovery rate
  • Procedure and different question
• Sequential hypothesis testing