Generalized Linear Models

Code: MA687 | L-T-P-C: 3-0-0-6

Prerequisites: MA589 (Statistical Foundations for Data Science) or MA324 (Statistical Inference and Multivariate Analysis) or Equivalent

Preamble / Objectives (Optional): Generalized linear model (GLM) is widely used in both academia and industry. In this course, the student will learn the theory related to GLM as well as when and how to apply GLMs in real data. We will cover GLMs for different types of response variables like continuous, binary, multinomial, and count. Topics will be covered with supporting data analysis in statistical software R.

Course Content/ Syllabus:Review of linear models: simple and multinomial linear regressions. Generalized linear models (GLM): the process of model fitting, components of GLM, measuring the goodness of fit, residual analysis, algorithm to fit GLMs.Models for binary data: logistic regressions, probit, log-log and others. Models for multinomial data: multinomial logit. Models for count data: Poisson, negative binomial and others, analysis of contingency table, analysis of ordinal data.

Text/References:

1. Peter McCullagh and John Nelder, Generalized Linear Models, Second Edition, Chapman and Hall/CRC (1989).
2. Julian J. Faraway, Extending the Linear Model with R: Generalized Linear, Mixed Effects, and Non-parametric Regression Models, Second Edition, Boca Raton: CRC Press (2006).