Please enroll for group one OR two!!!
Learning and qualification aims:
The students can describe, explain and systematise different advanced statistical and mathematical approaches to the quantitative analysis of geo- and environmental data and the modelling of human-environment systems, e.g. methods of applied and multivariate statistics, mathematical modelling and time series analysis. On the basis of the acquired theoretical and exemplified knowledge, the students can apply existing approaches independently and adapt them to specific problems where necessary. They can develop scientific research questions in the fields of data analysis and modelling and, using the acquired applied programming skills, plan and implement their own analyses.
Modulabschlussprüfung: Project work with programming elements connected to current research at the Institute. The project report will be written in form of a scientific article and handed in together with the programming code.
- Introduction to environmental modelling
- Mathematical preliminaries
- Parameter estimation & linear regression
- ANCOVA, multiple linear regression, dummy coding, collinearity, over-parameterisation, model comparison
- Generalised Linear Models (logistic & log-linear)
- Principle Component Analysis (PCA), Multivariate ANOVA (MANOVA), Discriminant Function Analysis (DFA)
- Measures of accuracy, confusion matrix, ROC/AUC, cross-validation; cluster analysis (kmeans & hierarchical)
- Introduction to spatial statistics
- Spatial autocorrelation
- Spatial weights and linear modelling
The seminar accompanies the lecture by Prof. Dr. Tobias Krüger and Prof. Dr. Tobia Lakes. We will apply the methods taught in the lecture using the open source programming language R (http://www.r-project.org/) and thus learn the basics concepts of scientific programming, advanced statistics and applied modelling. There will be homework. We expect the students to be familiar with the basic concepts of descriptive and test statistics.
[Dormann, C. (2013). Parametrische Statistik: Verteilungen, maximum likelihood und GLM in R. Springer. (German).]
Bolker B. (2008). Ecological Models and Data in R. Princeton University Press.
Zuur, A. (2007). Analyzing Ecological Data. Springer.