Generalized Linear Models

What it is about

You learn methods of regression for non-normally distributed target variables (e.g., binary, count data, nominal, positive values). In addition to classical GLMs such as logistic, Probit-, Poisson-, Gamma-, and log-linear models, extensions (e.g., overdispersion, random effects) are covered. By the end you will be able to estimate models, validate them, and analyze and interpret the results with R.

What you will be able to do

• Apply modern regression methods for non-normally distributed data
• Perform exploratory data analysis for regression problems
• Adapt and validate GLMs (including overdispersion, random effects)
• Interpretation of model results
• Analysis and interpretation with the statistics software R

What the module consists of

• VorlesungDelivery of content through lectures and example applications; motivation for independent literature work
• Übung/PraktikumPractical sessions with problem sheets and solutions for deepening and self-check

Teaching method

• VortragPresentation of concepts and demonstrative examples
• DiskussionDiscussion of the content together with the students
• Übungsaufgaben mit LösungenDeepening and self-control of what has been learned
• Eigenständige LiteraturarbeitEncouragement for independent analysis of the topics presented