For a long time, the working title of this book was Black Box Methods in SAS for Time Series Analysis. But in order to make the book look more serious, I changed the title to Practical Time Series Analysis Using SAS. The basic idea, however, has continued to be to consider SAS as a black box when extracting information such as forecasts from time series without specifying statistical models. This is important, as everyone knows that time series analysis is a complicated task for non-statisticians. This book is written for anyone who needs to generate seasonally adjusted series or forecasts, or to extract simple knowledge from observed series without too much effort.
This book also attempts to open the black box a bit to see what is inside. I am sure that some readers might want to close this black box as soon as possible and go straight to the detailed description of how to perform practical analyses. But other readers might want to have an idea of what is really going on behind all the PROCs. As such, I have used earlier versions of the text in courses on both theory and practical analyses for students of economics at the bachelor and master level at the University of Copenhagen.
Many years ago I wrote a textbook (in Danish) on time series analysis using the Box-Jenkins approach with applications of PROC ARIMA. It was in the day of punched cards. Computer time was an issue so users had to specify all models carefully, looking at autocorrelations, etc. Nowadays, this is superfluous; PROC VARMAX does that job in a split second. At that time, simple forecasting methods were doubtful, but with the improved flexibility offered by PROC ESM, they are now best practice. Moreover, the Kalman filter at that time was completely inaccessible for non-engineers. Now everyone can gain useful insight without even knowing that the Kalman filter has been applied. PROC UCM simply provides an intuitive way for coding, and the results are presented in impressive graphs.
All these user-friendly possibilities! I want to declare myself a great admirer of the ongoing development of the statistical facilities of SAS software.
Software
The code presented in this book is fully compatible with SAS 9.3 and with SAS/STAT 12.1.
Author Page
The example programs and data sets discussed in this book can be found at my author page: http://support.sas.com/publishing/authors/milhoj.html.
Anders Milhøj
March 2013