Probabilistic Modeling (NOT Performance Evaluation)

(There has been a name change planned for this course for a long time. Somehow it hasn't been processed yet.)

Professor: Norm Matloff
**Title:** Probablistic Modeling

**Prerequisites:** Calculus-based undergraduate course
in probability modeling, such as ECS 132 or STA 131A; knowledge of
matrix algebra, as in MAT 22A.

We will program in R. It is NOT assumed that you know R before taking the course; you will pick it up on your own if you don't have this background. Appendix A in the textbook gives a quick introduction to R.

**Grading:** Grade based mainly on
homework, but there will also be two quizzes.

**Topical Coverage:** Markov chains; random vectors and
matrix-valued analysis; mixture models, including EM algorithm;
prediction models (regression, classification, machine learning).

**Textbook:** N. Matloff,
*
From Algorithms to Z Scores: Probablistic and
Statistical Modeling in Computer Science
*
(open source, available
online),
covering parts of Chapters 6, 9, 11-13, 18, 19, 21-24

** Workload:**

- Work in teams of 3 or 4 students per group.
- Some math exercises, but the assignments will mainly consist of
writing quality R packages (3 or 4 over the time of the quarter).
**These WILL require mathematical analysis.** - Grading is interactive, by group.

Go here for further details.