Introduction to Generalized Linear Models with R

Date: Mon, Nov 24 - Fri, Nov 28 2014

Hour: 09:30

Speakers: Dae-Jin Lee, BCAM

Dates: 24/11/2014 - 28/11/2014 (5 sessions)
Time: 9:30 - 11:30 (a total of 10 hours)
Registration: CLOSED. Last update 3/11/2014: Due to the high demand on this course, we open a second edition of the course. The tentative dates will be 2nd-6th February 2015. Please fill the  form to register.

This course is oriented to postgraduate students, researchers and data analysts who need to move beyond standard linear models for modeling data that are not normally distributed. This short course provides an overview of generalized linear models (GLM's) using the R software. GLM's are most commonly used to model binary or count data. These types of data are very common in many research areas such as Biology, Medicine, Engineering, Business, Economics, and many other fields. In this course we will focus on real applications and examples with an emphasis on model validation, estimation and interpretation of the parameters and variable selection and goodness-of-fit.

You must bring your own laptop with R software installed, basic knowledge of R and linear models.


1. Introduction to Generalized Linear Models (GLM's)
1.1 Short review of linear regression.
1.2 What is a GLM and why to use them?
1.3 Components of a GLM
2. Models for binary data
2.1 Logistic regression
2.2 Estimation and interpretation of the parameters
2.3 Some examples
3. Multinomial regression
3.1 Function multinom() in R
3.2 Interpretation of the parameters
3.3 Variable selection
3.4 Some examples
4. Ordinal regression
4.1 Proportional odds model
4.2 Function polr() in R
4.3 Some examples
5.1 Poisson distribution
5.2 Poisson regression for incidence rates
5.3 Some examples



Confirmed speakers:

Dae-Jin Lee, BCAM