23313a
Vorlesung
SoSe 18: V Advanced Statistical Applications: from LM to GLMM using R
Alexandre Courtiol
Hinweise für Studierende
Zusätzl. Angaben / Voraussetzungen
Ersatztermin für den 1.5. und 10.5. der jeweilige Mittwoch;
open for Master students who took part in the course Introduction to R for statistical applications / Einführung in R für statistische Anwendungen or equivalent courses and open for PhD students (as far as there are places left by masterstudents); Schließen
open for Master students who took part in the course Introduction to R for statistical applications / Einführung in R für statistische Anwendungen or equivalent courses and open for PhD students (as far as there are places left by masterstudents); Schließen
Kommentar
GLM are a family of statistical models that aim at describing the effect of different variables (continuous and/or categorical) on one outcome of interest (continuous or categorical). GLM are widely used statistical tools needed by most biologists. By the end of the course, students will be able to perform a wide range of Generalized Linear Models (i.e. LM, GLM, LMM, GLMM) and understand more clearly the theory behind them. They will know how to translate biological problems into a GLM using R, identify the conditions where its application is appropriate, and draw inferences about biological systems from the outputs of the model. The content of the course is:
A . Linear models
- Introduction (response, predictors, parameters, design matrix, simulation)
- Point estimates (estimates, predictions, residuals)
- Uncertainty in point estimates (confidence intervals, prediction intervals, parametric bootstraps, likelihood profiling)
- Tests (t-test, F-test, LRT, multiple testing)
- Assumptions and Outliers (assumptions, checks, Box-Cox, outliers)
- Practice
B . Generalized Linear Models
- Introduction (link functions, variance functions, linear predictors, fitting procedure)
- Intervals & Tests (confidence intervals, prediction intervals, LRT)
- Residuals & Assumptions (assumptions, checks, types of residuals, GAM, Overdispersion & Zero-inflation, separation)
- Practice
C. Linear Mixed-effects Models
- Introduction (random effects, fitting procedure, predictions, BLUPs, )
- Special applications (quantitative genetics, phylogenetic regressions, meta-analyses)
- More complex mixed models (temporal and spatial autocorrelation, GLMM, non-gaussian random effects, HGLM, DHGLM, modeling heteroscedasticity) Schließen
A . Linear models
- Introduction (response, predictors, parameters, design matrix, simulation)
- Point estimates (estimates, predictions, residuals)
- Uncertainty in point estimates (confidence intervals, prediction intervals, parametric bootstraps, likelihood profiling)
- Tests (t-test, F-test, LRT, multiple testing)
- Assumptions and Outliers (assumptions, checks, Box-Cox, outliers)
- Practice
B . Generalized Linear Models
- Introduction (link functions, variance functions, linear predictors, fitting procedure)
- Intervals & Tests (confidence intervals, prediction intervals, LRT)
- Residuals & Assumptions (assumptions, checks, types of residuals, GAM, Overdispersion & Zero-inflation, separation)
- Practice
C. Linear Mixed-effects Models
- Introduction (random effects, fitting procedure, predictions, BLUPs, )
- Special applications (quantitative genetics, phylogenetic regressions, meta-analyses)
- More complex mixed models (temporal and spatial autocorrelation, GLMM, non-gaussian random effects, HGLM, DHGLM, modeling heteroscedasticity) Schließen