STATISTICS 926, Fall Semester 2014
Course Web Page --- an evolving document
Cumulative Course Materials:
Syllabus STAT 926
What this course is and is not:
- Purpose: This course is an extension of STAT 541.
- Prerequisites: Stat 541, the R language, linear algebra,
familiarity with Hilbert spaces.
- Like STAT 541, this course has
only one goal -- prepare statistics Ph.D. students for
- This is not an applied statistics course.
For a graduate level applied statistic course, see Paul
Rosenbaum's STAT 500 course on statistical methods.
- This course will be eclectic and make use of mathematics,
statistics, and R programming at all levels of difficulty.
- Andreas Buja
- Email: stat926.at.wharton[at-sign]gmail.com (urgent: buja.at.wharton[at-sign]gmail.com)
- Office hours: Fridays by appointment
- Office: JMHH 471
- Class Time: Mon and Wed, 12noon-1:30pm
- Class Room: JMHH F94
Tentative List of Topics (not necessarily in this order):
- Homework: Probably a half dozen, to be made up as we go along.
Homeworks will be the heart of what you retain from this course.
- Grades will be computed from homeworks and class participation alone.
- There will be no midterm or final exams.
- Visualization for data with many variables
- Classical multivariate methods based on eigen decompositions
- Review of the various classes of nonparametric regressions methods
which will be used as building blocks in nonparametric extensions of multivariate analysis
- Transformational multivariate analysis and scaling/scoring-based methods
- Functional multivariate analysis
- K-means clustering and its connection with principal components
- Multidimensional scaling, manifold learning, graph drawing.
- Writing for research, including style and clarity, typesetting, web publishing
Tentative recommended texts:
- As the tool of choice for the implementation of methods
and for simulations, we use the
R programming language.
- Note: This is not an R class. R will be assumed and not
As we go along, further special topics books and articles may be recommended.
- Deepayan Sarkar on "Lattice -- Multivariate Data Visualization with R"
- Hastie, Tibshirani, Friedman on "Elements of Statistical Learning"
- Alan Izenman on "Modern Multivariate Statistical Techniques"