QERM 514/Presentations

From QERM Wiki

Presentation schedule: Please try to fill up one presentation per day. If you really want to do one topic particularly, and some one has already signed up, we can do two on one day. I (that is, Hans) encourage you to sign up for topics which are (i) interesting to you and (ii) something you want to do some reading on. You'll get the most out of it if you challenge yourself.

Remember there is a written component as well as an in-class presentation. The in-class presentation is on the day the material is presented--so you will have to read ahead a bit for that topic. Derek and I would be happy to help you find an example for your topic if you don't have one in mind (as well as find additional reading if you think it would be beneficial).

• Linear predictors - April 8 (Open)

Fitting a line to continuous predictors, calculating R², using matrix notation for lines. This is ordinary linear regression, maybe some transformations of predictors.

• Linear inference - April 10 (Open)

Testing a hypothesis, both F and t tests, developing confidence intervals, prediction confidence intervals.

• One way anova - April 15 (Open)

Models with a continuous response, normal error and categorical predictors, coding and contrasts.

• Two-way anova - April 17 (Open)

Two or more categorical predictors, interactions, interaction plots.

• Ancova - April 22 (Amy)

Mixed continuous and categorical predictors.

• Variance stabilizing transformations - April 24 (Liz Atwood)

Deviations from the normal error assumptions.

• Outliers - April 29 (Chloe Bracis)

Identifying outliers or influential observations using a variety of techniques.

• Modeling practice - May 1 (Tommy Garrison)

Putting together ideas; anything from previous lectures (or interesting extensions).

• Maximum likelihood methods - May 6 (Open)

Fitting statistical models using general optimization techniques.

• Nonlinear least squares - May 8 (Motoki Wu)

Fitting models using nonlinear least squares, or interpreting the results.

• Generalized linear models 1 - May 13 (Open)

Binomial and Poisson regression, link functions, deviance, anodev

• Generalized linear models 2 - May 15 (Joe Anderson)

Checking model adequacy, overdispersion, other glms.

• GLM Practice - May 20 (Open)

Anything in the glm framework.

• Linear mixed effects - May 22 (Open)

Modeling hierarchical data.

• Model selection - May 27 (Open)

Backwards selection, multiple comparisons, AIC and AICc

• Extensions of linear models - May 29 (Open)

Terminology and ideas of other techniques related to linear models.