Lecture 6

17 Abril 2020, 08:00 • RUI MIGUEL BATISTA PAULO

Inference in the context of a GLM: the iterative rewheighted least squares algorithm for estimating the coefficients of the linear predictor.

Estimation of the dispersion parameter.

Pearson and deviance residuals.

Solution of question 2 of exercise sheet 1.

---

Lecture 5

3 Abril 2020, 08:00 • RUI MIGUEL BATISTA PAULO

Video-lecture.

Saturated model, likehood ratio between current and saturated model.

Deviance and scaled deviance. Assessing the statistical significance of nested models. Table of deviances. Examples.

The normal case: the scaled deviance is the residual sum of squares. Relationship with the ANOVA F-tests.

---

Lecture 4

27 Março 2020, 08:00 • RUI MIGUEL BATISTA PAULO

Video-lecture.

Generalized linear models: distribution for the data, linear predictor, link function. Examples.

Link function: remarks, canonical link function.

Example: binary response, canonical link function, the SwissLabor dataset. Factor as a regressor. Dummy variables, interpretation of the parameters as increments in the intercept over the reference level. Interaction terms as increments in the slope over the reference level. 

Factors with r levels: number of parameters associated. Interaction terms: number of parameters associated.

---

Lecture 3

20 Março 2020, 08:00 • RUI MIGUEL BATISTA PAULO

Video-lecture available from TEAMS.

Multiple linear regression: F-tests and diagnostic plots.

Generalized linear models: introduction. The exponential family of distributions: definition, natural parameter, mean and variance. Variance function. Examples.

---

Classes suspended

13 Março 2020, 08:00 • RUI MIGUEL BATISTA PAULO

No class due to the covid-19 outbreak.

---