STA 410/2102 Statistical Computation
- When and where: Mondays 9-12pm in SS 1086
- Radu's office hours: Monday 3-4pm in SS 6010
- TA: Gun Ho Jang: Office hours: Friday 3-4pm (starts Sept 19) in SS 2128; email: smsroot1 AT gmail DOT com.
-
- Email policy: Please try to talk to me after class or during office hours. My schedule does not allow me to answer
questions by email.
- The Midterm Exam is scheduled for November 17, during class. Material presented in class until November 11 is covered
by the Midterm.
- The Final Exam is scheduled for December 10, 9am-12pm in BN3.
- Some old tests are available here, here, here and here. Note that this year I decided for the first time to add the EM algorithm to the list of topics.
- DO NOT FORGET TO BRING A CALCULATOR
- Homework 2 has been posted below. It is due on Dec 1, in class.
- Gun Ho's office hours will be held in the last two weeks of the term in SS2111.
- This course studies various computational methods used primarily in statistical problems. I also would like to
develop students' abilities to write programs for statistical problems
that are not handled by standard packages. Students will learn the
capabilities of the R statistical software and learn to
program new statistical methods in that environment. R will be
introduced as part of the course; no prior knowledge of it is
necessary.
- Prerequisites: This course is designed for graduate and
senior undergraduate students in statistics, actuarial science,
computer science or other fields where statistical computation is
important. Students should have a basic background in statistical
methods (eg. at the level of STA302), and some prior experience with
programming (eg. at the level of CSC108).
- Textbook: There is no textbook that fits this course. The main source
for preparing the exams are the class notes. Please attend all classes.
- Marking:
30% based on one in-class 2-hour midterm
50% based on final exam
20% based on homework
- You can download R for Windows freely at http://lib.stat.cmu.edu/R/CRAN/bin/windows/base/ (you need to click on the file with suffix .exe).
Tentative Syllabus:
- Topic 1: Data manipulation in R
- Topic 2: Linear algebra: manipulations of vectors and matrices
- Topic 3: Singular value decomposition and principal components
- Topic 4: Weighted linear regression
- Topic 5: Numerical Optimisation Methods: The EM algorithm and Newton-Raphson
- Topic 6: Generalized linear models and Fisher Scoring
- Topic 7: Monte Carlo methods
- Topic 8: Bootstrap
- Topic 9: Density estimation
- Topic 10: Cross-validation
Handouts
- This is a brief introduction to R .
- The commands used for the egg example discussed in the second class are
here . The data are here .
- A very brief outline of what was covered in the first two classes is
available here
- Student score example for PCA is available here . The student score data is here .
- The regression handout is here.
The datasets used are:
- There are two handouts for the EM algorithm. The EM for a simple regression model and the EM for the blood type example
- The Newton-Raphson handout is here
- The Monte Carlo handout is here
- First bootstrap handout is here . The bootstrap for regression handout is
here.
Homeworks
- Homework number 1 is here. The file mixedata.txt is
here. .
- Homework number 2 is here.
|