Purpose
This course is on the X-12-ARIMA seasonal adjustment method to estimate the trend-cycle, seasonal, holiday, trading-day and irregular components of a time series. The course includes theory and demonstrations of the seasonal adjustment softwares. The purpose of this course is:
- to understand the components of time series;
- to get familiar with the options and statistical methods used in X-12-ARIMA;
- to learn how to assess the quality of the seasonal adjustment results;
- to become familiar with the X-12-ARIMA software and/or the interface to run it;
- to become familiar with X13graphjava tool to produce specialized graphs related to the time series components.
Benefits to participants
Upon completion of the course, the participants will be more familiar with many options in the X-12-ARIMA program and therefore will be able to better assess the quality of a seasonal adjustment. They will see through demonstrations how to apply the methods to select options for the X-12-ARIMA method and assess the obtained results. The course is theoretical and technical.
Target population
This course is intended for employees involved or interested in the production and analysis of seasonally adjusted series.
Course outline
The course examines
- the components of time series (summary);
- the calculations done in the method, such as moving averages and treatment of extreme observations;
- the choice of the decomposition model;
- the ARIMA forecasting as part of the X-12-ARIMA method;
- the estimation of calendar effects such as the Easter and trading day effects and the treatment of outliers by ARIMA regression;
- the direct versus the indirect seasonal adjustment;
- the tests used to assess the results of the seasonal adjustment;
- the overall strategy and criteria which should be used to do seasonal adjustment;
- and how to do all of the above specifically with the X-12-ARIMA program.
Other Related Courses
The course is specialized and requires basic statistical knowledge. The course STC0431. The components of Time Series should be taken first.
Delivery type: Virtual instructor-led
Duration: 3 half-days
Contact:
If you have questions or to register to the course, contact us at statcan.timeseriessupportsoutienenserieschronologiques.statcan@statcan.gc.ca