|Page tools: Print Page Print All|
Time Series Methods Course Revision
Time Series training for methodologists has been substantially revised this year. The course aims to provide ABS methodologists with both theoretical and practical experience with seasonal time series, with particular emphasis on the real world issues and changes that require methodologists' attention and intervention. There were major updates to the course notes, and the course was delivered in a two week intensive format. The course consists of a set of lectures, course notes, tutorials, assignments and an exam.
The course covers: Time series decomposition; interpreting time series; Fourier analysis; filters; gain functions; prior corrections, including moving holidays and trading day corrections; ARIMA models; regression-ARIMA; and backcasting (to handle classification changes).
Some more new sections will be written and the course notes as a whole reviewed before re-releasing the notes on the ABS website during 2013.
These documents will be presented in a new window.