620-374: Time Series and Forecasting Section
Notes
Additional ARMA notes by Rothenberg
Data for the exercises is taken from the book of Makridakis, Wheelwright and Hyndman, and can also be found at
http://www-personal.buseco.monash.edu.au/~hyndman/forecasting/
Solutions to selected questions:
Lab Tasks
- Lab 1 Task is to fit a 7-MA and a centered 12-MA smoother to the US housing sales time-series.
Once you have done that, use the detrended series to estimate the additive seasonal component.
Solution in Excel and in R
- Lab 2 Task is Question 5 from the exercise sheet.
Solution in Excel and in R
- Lab 3 Task is Question 17 from the exercise sheet.
The random numbers are given as random.csv above.
You might like to repeat the exercise with a longer sequence of random numbers (for example, use rnorm(100) in R)
Series calculated using Excel and acf/pacf plots with commentary
Assignments
There will be three assignments for this section of the course, due on Fridays at 5:00pm on weeks 6, 7 and 8.
Each will be worth 20/9%.
Assignments can be dropped into the assignment box on the ground floor of the Richard Berry building, or else handed to the lecturer in person.
Assignments up to a week late receive only half marks.
Assignments over a week late will receive no marks, but will still be marked if you wish feedback.
Make sure you put the course code and your name clearly on the top of each assignment. Collaboration on assignments is not allowed.
- Assignment 1 is Exercise Question 4
- Assignment 2 is Exercise Question 8
- Assignment 3 is Exercise Questions 13 and 14
References
A good text for this part of the course is
- Chris Chatfield, The Analysis of Time Series: An Introduction, 6th Edition. Chapman and Hall/CRC, 2004.
A less technical but useful text is
- Spyros Makridakis, Steven C. Wheelwright and Rob J. Hyndman, Forecasting Methods and Applications, 3rd Edition. Wiley, 1998.