What is this course about?
As the objective of this course you will
- gain an overview on extended regression modelling techniques that allow to analyse data with non-normal responses,
- learn about approaches for modeling nonlinear effects in scatterplot smoothing,
- get an introduction to additive models for complex regression analyses, and
- learn how to implement these approaches using statistical software packages.
Generalized linear models (binary and Poisson regression, exponential families, maximum likelihood estimation, iteratively weighted least squares regression, tests of hypotheses, confidence intervals, model selection and model checking, categorical regression models), nonparametric smoothing techniques (penalized spline smoothing, local smoothing approaches, general properties of scatterplot smoothers, choosing the smoothing parameter, bivariate and spatial smoothing, generalized additive models).
In the exam, the students demonstrate their ability to choose, fit and interpret extended regression modeling techniques. They show a general understanding of the derived estimates and their interpretation in various contexts. The students are able to implement complex regression models using statistical software and to interpret the corresponding results. The exam covers contents of both the lecture and the exercise class.
Where? When? Who?
Lectures: MZG 8.163, Thursdays from 14 to 16, Prof. Dr. Thomas Kneib
Exercises: MZG 5.111, Wednesdays from 10 to 12, Holger Reulen