|
|
|
|
|
| Purpose |
To acquaint graduate students from various
disciplines with a firm understanding of the ARIMA(p,d,q) class of
models and to use this information to fit appropriate models to real
data and forecast if desired; to familiarize students with the
frequency domain approach including the definition, interpretation and
estimation of the spectral density. |
| Prerequisites |
STAT 704 and 512. Stochastic properties,
identification, estimation, and forecasting methods for stationary and
nonstationary time series models. Talk
to the
instructor to get a permission to register for the course. |
| Textbook |
Required: Introduction
to Time Series and Forecasting,
2nd edition, by P.J. Brockwell and R.A. Davis, Springer, 2002.
Recommended:
Time Series: Theory and Methods, 2nd edition, P.J. Brockwell and R.A.
Davis,
Springer,
1991.
|
| Grading |
Exam 1 (20%, in
class, open book),
Exam 2 (20%, in class, open book), Final Exam
(30%, take home), Project (20%) and homework problems
(10%). |
Grading (%):
A 90; B+ 87; B
80; C+ 77; C 70;
D+ 67; D 60; F
<60;
|
| Software |
ITSM2000 – comes free with the
textbook.
Specialized
software for time series analysis and forecasting.
User
friendly with pull down menus and good graphical capabilities. |
| Selected Topics |
1.
Box-Jenkins ARIMA
2. Seasonal ARIMA models
3. Unit roots
4. Cointegration
5. Granger causality
6. Vector Autoregression (VAR) models
7. Vector Error Correction (VEC) models
8. Intervention analysis and structural change
9. State-Space models |
10.
Markov regime-switching model
11.
ARCH models
12. GARCH models
13. Smoothing
14. Forecasting techniques
15. Simulations |
|
|
|