STATS305B: Methods for Applied Statistics

Syllabus

Instructor & TAs

Instructors

Jonathan Taylor

  • Office hours: Wednesday 1:30-3:30pm, Sequoia Hall #137. See Canvas page for zoom meetings during first two weeks.

Teaching Assistants

Piazza

Here is the piazza page

Gradescope

Here is the gradescope page

Schedule

MW 11:30-1:00, Building 200, Rm 303

Textbook

Lectures will be a mix between chalk talk and computing examples.

Additional comments

Computing environment

We will use R for most calculations. Examples will typically be in the form of jupyter notebooks.

Prerequisites

Stats 305A or similar.

Course description

The first half of the course will be from the textbook, focusing on logistic regression and log-linear models.

The second half of the course will consist of different topics, focusing on exponential families:

  • Regularized regression: LASSO / group LASSO for logistic and log-linear model

  • General properties of exponential families

  • Bayesian analysis / conjugate priors

  • EM algorithm

  • Pseudo-likelihood

  • Gibbs sampling

  • Sampling of topics in survival analysis: censoring / log-rank test / Cox proportional hazards model.

Evaluation

  • 5 assignments (70%)

  • final exam (30%) (according to Stanford calendar: T 03/15 @ 8:30AM)

Practice exam

  • A previous year’s exam available here

Canvas page

  • Find the canvas page here

Slides

Notes on these pages are available as HTML slides: