# Lectures in STK 3100/4100 fall 2019

Below are given plans for the coming lectures and an overview of the material that has been covered in the course so far. Section numbers, etc refer to the book by Alan Agresti.

**Previous lectures**

*Week 34 (August 19th and 21st):*

__Chapter 1__: Introduction to generalized linear models (section 1.1; except 1.1.6), model matrices and model spaces (section 1.3), identifiability of model parameters (section 1.4.1). Section 1.2 contains material that should be known from earlier courses, and it will not be discussed at the lectures. But the students should read section 1.2 on their own. (Slides, R code for data examples)

__Chapter 2__: Least squares model fitting (sections 2.1.1, 2.1.2, and 2.1.3). (Lecture, R code for illustration of model matrices)

*Week 35 (August 26th):*

__Chapter 2__: More on least squares model fitting (sections 2.1.4 and 2.1.5), projections of data onto model spaces (section 2.2, except 2.2.4), projections and decomposition of sums of squares for linear models (section 2.3 except 2.3.4). (Lecture)

*Week 36 (September 2nd):*

__Chapter 2__: Estimation of error variance (section 2.4.1) and sums of squares (2.4.2).

__Chapter 3__: Distribution theory for normal variates (section 3.1, only page 85 in 3.1.4 and except 3.1.5), hypothesis testing for one-way layout (section 3.2.1). (Lecture, R-commands)

*Week 37 (September 9th):*

__Chapter 3__: Hypothesis testing for linear normal models (sections 3.2.2 and 3.2.5).

__Chapters 4 and 5__: Logistic regression model (section 5.1) and exponential dispersion family for GLMs (section 4.1). (Lecture)

*Week 38 (September 16th):*

__Chapters 4 and 5__: Likelihood and approximate distributions for GLMs (section 4.2, except 4.2.5 and 4.2.6; section 4.5.5; sections 5.3.1 and 5.3.2). Hypothesis testing (section 4.3, except 4.3.5 and 4.3.6). Interpretation of parameters in logistic regression (section 5.2.1).

*Week 39 (September 23rd):*

__Chapters 4 and 5__: Confidence intervals (sections 4.3.5 and 4.3.6), delta method and fitted values (section 4.2.5), deviance and sum of squares (sections 4.4.1, 4.4.2). (Slides, R-code, we also look more carefully at R-code 2 from the previous lecture)

*Week 40 (September 30th):*

__Chapters 4 and 5__: Deviance and likelihood ratio testing (sections 4.4.1, 4.4.2, 4.4.3, 5.5.1, and 5.5.2), Pearson chi-square statistics (sections 4.4.4, 5.5.1, and 5.5.2), deviance and Pearson residuals (section 4.4.6), Newton-Raphson method (section 4.5.1), Fischer scoring and iteratively reweighted least squares (sections 4.5.2, 4.5.4, and 4.5.5). (Slides, Rcode)

*Week 41 (October 7th)*

__Chapter 4__: Selecting explanatory variables for a GLM (sections 4.6.1, 4.6.2, and 4.6.3), normal linear models and gamma GLMs (sections 4.7.1, 4.7.2, 4.7.3). Link functions for binomial data (sections 5.6.1, 5.6.3, and 5.7.2), various topics of GLMs for binomial data (sections 5.2.2, 5.2.4), (slides, R-code for model selection, gamma-regression and link functions).

*Week 42 (October 14th)*

__Chapters 6:__ Some remaining topics from Chapter 5 (5.2.3, 5.2.4, 5.25, 5.3.4, 5.4.2). Baseline-category logit models for multinomial responses (sections 6.1.1, 6.1.2, 6.1.3, 6.1.4). (Lecture, R-code for ROC-curves and multinomial logit)

*Week 43 (October 21st):*

__Chapters 7:__ Ordinal multinomial models (6.2.1, 6.2.2, 6.3.2 and 6.3.3), Poisson GLMs (sections 7.1.1, 7.1.3, 7.1.4, and 7.1.6). (Lecture, R-code for cumulative logit model on mental impairment data and for Poisson regression on lung cancer and insurance claims data)

*Week 44 (October 28th):*

__Chapters 7 and 8:__ Overdispersion and negative binomial GLMs (sections 7.3.1, 7.3.2, 7.3.3, and 7.3.4), zero-inflated GLMs (sections 7.4.1, 7.5.1, and 7.5.2), quasi-likelihood methods and variance inflation [sections 8.1.1, 8.1.2, 8.1.3]. (Lecture, R-code for Crab-data)

*Week 45 (November 4th):*

__Chapter 9:__ More on quasi-likelihood [8.3.1, 8.3.2, 8.3.3, and 8.3.4]. Normal linear mixed models and marginal linear models (sections 9.2, 9.3.1, 9.3.3, and 9.6.1). (Lecture, R-code for Crabs-data, Fecal fat, FEV-data and Smoking prevention).

*Week 46 (November 11th):*

__Chapter 9:__ Prediction of random effects for normal linear mixed models (section 9.3.2). Marginal and generalized linear mixed models (sections 9.1.2 and 9.1.3), binomial and Poisson GLMMs (sections 9.4.1, 9.4.2, and 9.7), ML-estimation for GLMMs (sections 9.5.1 and 9.5.2), marginal models (sections 9.6.3 and 9.6.4). (Lecture, R-code for prediction of variance component for Fecal fat data and FEV-data, R-code for GLMM/GEE example).

*Week 47 (November 18th):*

Recapitulation of the main points in the curriculum.

Exam 2017, Problems 1, 3 and 4.

**Plans for coming lectures**

*Week 48 (November 25th):*

Exam 2018, Problems 1 and 2.

Exam 2015, Problem 2 and 3

*Week 49 (December 2nd): *

Exam 2014, Problems 1, 2 and 3