Welcome to DSCI 562: Regression II#

This course focuses on exploring regression techniques beyond Ordinary Least-squares (OLS). Some specific inquiries are the following:

  • What if the response is still continuous but non-negative, or our observations are not independent anymore?

  • Maybe the response is binary, a count, or categorical.

  • Moreover, what if we have censored data?

  • In many other cases, we might be interested in modelling a response different from a conditioned mean. Different Data Science-related projects will require making inference on different conditioned response quantiles.

We cover useful extensions to ordinary linear regression: generalized linear models (GLMs), mixed-effects, local, survival, and quantile regression, and techniques for dealing with missing data.

High-Level Goals#

By the end of the course, students are expected to:

  • Describe the risk and value of making parametric assumptions in regression.

  • Fit model functions that represent probabilistic quantities besides the mean.

  • Identify situations where OLS regression is sub-optimal, and apply alternative regression methods that better address the situation.

Assessments#

This is an assignment-based course. The following deliverables will determine your course grade:

Assessment

Weight

Lab Assignment 1

12%

Lab Assignment 2

12%

Lab Assignment 3

12%

Lab Assignment 4

12%

Quiz 1

25%

Quiz 2

25%

Lecture Attendance (iClicker)

2%

Lecture Schedule#

This course occurs during Block 4 in the academic year. The course notes can be accessed here. Typically, students are expected to review these notes before each lecture.

Note: Answers to the iClicker questions are not included in these public materials.

Lecture

Topic

Suggested Reading Material

1

Generalized Linear Models: Link Functions and Count Regression

2

Generalized Linear Models: Model Selection and Multinomial Logistic Regression

3

Generalized Linear Models: Ordinal Logistic Regression

4

Linear Mixed-Effects Models

5

Survival Analysis

See Kleinbaum and Klein (2005) in Reference Material

6

Local Regression

  • ISL 7.2 Step Functions
  • ISL 7.6 Local Regression

7

Quantile Regression

  • Fahrmeir (2013) 10. Quantile Regression (except 10.2.2 Bayesian Quantile Regression)

8

Missing Data

See van Buuren, S. (2012) in Reference Material

See the lecture learning objectives for a detailed breakdown of lecture-by-lecture learning objectives.

Reference Material#

  • Agresti, A (2012). Categorical Data Analysis, John Wiley & Sons, Incorporated. ProQuest Ebook Central.

    • This book is helpful for generalized linear models with discrete responses.

  • Collett, D. (2003). Modelling Binary Data (2nd ed.). Chapman and Hall/CRC. https://doi.org/10.1201/b16654

  • Fahrmeir, L. (2013). Regression Models, Methods and Applications. Springer Berlin Heidelberg.

  • Faraway, Julian J. (2005). Extending the Linear Model with R: Generalized Linear, Mixed Effects and Nonparametric Regression Models, CRC Press LLC. ProQuest Ebook Central.

    • This book is great for learning how to work within the R environment with the models we will be working on. Its approach is essentially practical.

  • Gelman, A. and Hill, J. (2007). Data Analysis Using Regression and Multilevel/Hierarchical Models. Analytical Methods for Social Research. Cambridge University Press.

    • This book is pretty useful and practical as introductory material.

  • James, G., Witten, D., Hastie, T., and Tibshirani, R. (2014). An Introduction to Statistical Learning: with Applications in R. Springer Publishing Company, Incorporated.

  • Kleinbaum, D. G. and Klein, M. (2005). Survival analysis : A Self-Learning Text. Springer.

    • This book is a good start for Survival Analysis:

      • Chapter 1 (Introduction): Introduction to Survival Analysis (I), Censored Data (II), Terminology and Notation (III).

      • Chapter 2 (Kaplan-Meier Curves): Review (I), Example of Kaplan-Meier Curves (II), General Features of Kaplan-Meier Curves (III), Confidence Intervals for Kaplan-Meier Curves (VII and VIII).

      • Chapter 3 (Cox Proportional Hazards Model): Example of Cox Proportional Hazards Model (I), Formula of Cox Proportional Hazards Model (II), Why the Cox Proportional Hazards Model is Popular (III), Estimation of the Cox Proportional Hazards Model (IV).

      • Chapter 7 (Parametric Model): Overview (I), Relationship Between the Probability Density Function with Hazard and Survival Functions (II), Weibull Example (IV).

  • Roback, P. and Legler, J. (2020). Beyond Multiple Linear Regression.

  • Rubin, D. B. (1987). Multiple Imputation for Nonresponse in Surveys. Wiley.

  • van Buuren, S. (2012). Flexible Imputation of Missing Data:

    • 1.1 The problem of missing data.

    • 1.2 Concepts of MCAR, MAR and MNAR.

    • 1.3 Ad-hoc solutions.

    • 1.4 Multiple imputation in a nutshell.

Dataset References#

These are the papers from which each dataset used in the lectures comes from:

Policies#

See the general MDS policies.

Previous Release#

A 2019 release of these course’s notes can be found here.

Attribution#

The course is built upon previous years’ materials developed by previous instructors.

License#

See our license for further details.

© 2024 G. Alexi Rodríguez-Arelis, Rodolfo Lourenzutti, and Vincenzo Coia