Building on the statistical inference concepts (likelihood and Bayes) from Statistical Inference I, this class will cover more advanced topics relevant in  contemporary  ('computer-age') statistical inference including

  • topics (in particular) relevant in high-dimensional setting: Multiple testing, inference after model selection, reproducibility,
  • regularization (Ridge, Lasso) and connections to Bayesian inference
  • inference in more complex settings: model misspecification, dependent data, missing data, censored data
  • computationally intensive methods: More on Bayesian inference, on  the bootstrap and resampling procedures, permutation tests

We will often discuss frequentist and Bayesian approaches to the same problem as well as connections between them.


Semester: SoSe 2022