Stat 637: Generalized Linear Models
Content for this course will be available on the course Learning Suite webpage.
Course Description and Learning Outcomes
This course aims to introduce the statistical theory to extend regression and analysis of variance to non-normal data.
- Fit and choose an appropriate generalized linear model for binary, ordered categorical, un-ordered categorical, and count response variables using R and SAS.
- Evaluate the validity/appropriateness of the chosen model using model diagnostics such as residual plots and deviance.
- Identify weaknesses in the chosen model for a given data set.
- Identify when overdispersion is present in a given data set and ways to account for overdispersion in the model.
- Make predictions and determine confidence intervals using the fitted model.
- Demonstrate understanding of the connection between Normal linear model theory and generalized linear model theory by expressing the Normal linear model as a generalized linear model.
- Determine the canonical link for any distribution in the exponential family.
- Reproduce (for any distribution in the exponential family) score equations, Fisher information, and write out the form of the iterative reweighted least squares algorithm for finding maximum likelihood estimates of the coefficients.
- Time permitting, implement a simple multi-level model using R and SAS and a Bayesian logistic regression model using R.