Syllabus of STAT 685, Summer 2007
H. Joseph Newton
Professor of Statistics and Dean of Science
Texas A&M University
Class: T-Th, 1:30-3:00, Blocker 517
Office: 517 Blocker Building.
Hours: Daily, 1:00-2:00, or by
appointment.
Phone: (979) 845-8817.
FAX: (979) 845-6077.
E-mail:
jnewton@stat.tamu.edu
Address:
Statistics Department ,
Texas A&M University ,
College Station TX 77843-3143 USA.
Prerequisites
Knowledge of random variables, moments, distribution theory, maximum
likelihood estimation, and regression analysis (using matrices) such
as covered in STAT 601 is assumed; complex numbers and trigonometry.
Course Materials
The text for the course is a reprint of ``TIMESLAB: A Time Series
Analysis Laboratory,'' written by the instructor and originally
published by Wadsworth & Brooks/Cole. Try to get copies from people who
took the course before.
A set of course notes is available on the course web page under
``Topic pdf files'' on
the ``Course notes'' page.
The software for the course is a translation of the timeslab software that accompanies the
text into the freely available R language (R is available for free download to
PCs)
Determining the Course Grade
-
Two in-class, closed-book exams (worth 30\% each) and a final
project worth 40%;
Homework (primarily analyzing real time series) worth total of 40%.
-
We will not have a final exam.
-
The last day of class will be determined (early in August).
Course Outline
- Lecture 1
Introduction; Correlogram, Partial Correlogram
(Section 1.1-1.4.2)
- Lecture 2: Introduction to the TIMESLAB program; Periodogram
(Appendix B, Section 1.4)
- Lecture 3: Periodogram (continued); Transforming data
(Section 1.5)
- Lecture 4: Simple forecasting methods (Section 1.6)
- Lecture 5: Difference equations (Section 1.6)
- Lecture 6: Review for exam 1
- Lecture 7: Exam 1: Covers Chapter 1
- Lecture 8: Covariance stationary time series
(Sections 2.1-2.2)
- Lecture 9: Linear filters (Section 2.3)
- Lecture 10: Theory of prediction (Section 2.4)
- Lecture 11: Random walks; ARMA processes (Section 2.5)
- Lecture 12: ARMA processes (continued)
- Lecture 13: Statistical properties of descriptive
statistics (Section 3.1)
- Lecture 14: Tests for white noise (Section 3.2)
- Lecture 15: Smooth periodogram spectral estimation (Section 3.3)
- Lecture 16: Review for exam 2
- Lecture 17: Exam 2: Covers through Section 3.3
- Lecture 18: Fitting ARMA models (Sections 3.4-3.5)
- Lecture 19: Fitting ARMA models; Box-Jenkins
forecasting (Section 3.6)
- Lecture 20: Box-Jenkins forecasting and other modeling
strategies (Sections 3.6-3.7)
- Lecture 21: Searching for periodicities (Sections 3.8-3.9)
- Lecture 22: Bivariate time series (Section 4.1)
- Lecture 23: Coherence, phase, and gain (Sections 4.1-4.2)
- Lecture 24: Additional topics
- Lecture 25: Last class; Final project reports