Usage Note 56928: Data pre-processing requirements for SAS® Forecast Studio
 |  |  |  |
Data is often not in the appropriate condition to be read into SAS Forecast Studio. To avoid misleading or incorrect analyses, pre-process your data before creating projects in SAS Forecast Studio.
The following list outlines requirements, recommendations, and potential consequences regarding pre-processing your data.
Requirements:
- Accumulate your data according to a certain frequency. Examples: hourly, daily, weekly, monthly, or yearly. The method for accumulating observations within each time period is based on a particular statistic. Examples: sum, mean, median, minimum, maximum, standard deviation, and so on. Choose the appropriate accumulation frequency that is relevant for your data.
- Equally space your data so that successive observations are a fixed time interval apart. Examples: hourly, daily, weekly, monthly, or yearly.
- Remove duplicate observations.
- Check inactive parent series for any new or existing child series that have new data.
- Remove retired series.
- Remove series that have few or no observations in the in-sample region. If series have little or no data, then SAS Forecast Studio might fail to produce forecasts and models. Perform the same action for the series in the holdout and out-of-sample regions.
- For each independent variable, make sure that variability exists in the in-sample region. Example: make sure that the independent variable does not consist of only zeros.
- For each series, make sure that the holdout is adequate and does not consist of only zeros or missing values.
Recommendations:
- Try grouping similar types of series into one project before reading them into SAS Forecast Studio. Example: group short series and long series in separate projects rather than combining them in one project. There is no consistent definition of what constitutes a short series.
- If you are using macros such as %FSCREATE, then try creating the project through the SAS Forecast Studio user interface. The user interface has a more stringent time-interval check.
Potential consequences of not pre-processing your data:
- The project is not created.
- The time interval cannot be determined.
- A "Model List Problem" error occurs. This error does not affect reconciliation. The error indicates that the forecast of the selected model failed, and that instead the model ESMBEST is used to generate a forecast.
- A "Failed to Forecast" error occurs. This error does affect reconciliation because a series reconciliation is influenced by its siblings' failed forecasts. The error might be caused by having little or no data.
Prepare projects in a way that reduces the number of failed forecasts to the smallest possible amount. This reduction can be achieved by minimizing the number of "Model List Problem" errors and by completely eliminating the number of "Failed to Forecast" errors.
Even after pre-processing your data, there might be issues that occur in SAS Forecast Studio projects due to a lack of adequate data. If issues are still present after your data has been pre-processed, then determine whether the issues are affected by holdouts, model selection lists, and so on.
For code-based examples that are related to accumulating and removing duplicates from your data, reference SAS Note 54882: "The SAS Forecast Studio New Project Creation wizard gives an "Unable to detect time interval ..." warning." For general requirements and best practices that are related to SAS Forecast Server usage, reference SAS Note 54702: "Requirements and best practices for SAS® Forecast Server."
Operating System and Release Information
| Product Family | Product | System | SAS Release | |
| Reported | Fixed* | |||
| SAS System | SAS Forecast Studio | Microsoft® Windows® for x64 | ||
| Microsoft Windows 8 Enterprise 32-bit | ||||
| Microsoft Windows 8 Enterprise x64 | ||||
| Microsoft Windows 8 Pro 32-bit | ||||
| Microsoft Windows 8 Pro x64 | ||||
| Microsoft Windows 8.1 Enterprise 32-bit | ||||
| Microsoft Windows 8.1 Enterprise x64 | ||||
| Microsoft Windows 8.1 Pro | ||||
| Microsoft Windows 8.1 Pro 32-bit | ||||
| Microsoft Windows 95/98 | ||||
| Microsoft Windows 2000 Advanced Server | ||||
| Microsoft Windows 2000 Datacenter Server | ||||
| Microsoft Windows 2000 Server | ||||
| Microsoft Windows 2000 Professional | ||||
| Microsoft Windows NT Workstation | ||||
| Microsoft Windows Server 2003 Datacenter Edition | ||||
| Microsoft Windows Server 2003 Enterprise Edition | ||||
| Microsoft Windows Server 2003 Standard Edition | ||||
| Microsoft Windows Server 2003 for x64 | ||||
| Microsoft Windows Server 2008 | ||||
| Microsoft Windows Server 2008 R2 | ||||
| Microsoft Windows Server 2008 for x64 | ||||
| Microsoft Windows Server 2012 Datacenter | ||||
| Microsoft Windows Server 2012 R2 Datacenter | ||||
| Microsoft Windows Server 2012 R2 Std | ||||
| Microsoft Windows Server 2012 Std | ||||
| Microsoft Windows XP Professional | ||||
| Windows 7 Enterprise 32 bit | ||||
| Windows 7 Enterprise x64 | ||||
| Windows 7 Home Premium 32 bit | ||||
| Windows 7 Home Premium x64 | ||||
| Windows 7 Professional 32 bit | ||||
| Windows 7 Professional x64 | ||||
| Windows 7 Ultimate 32 bit | ||||
| Windows 7 Ultimate x64 | ||||
| Windows Millennium Edition (Me) | ||||
| Windows Vista | ||||
| Windows Vista for x64 | ||||
| Type: | Usage Note |
| Priority: |
| Date Modified: | 2016-02-22 08:43:10 |
| Date Created: | 2015-10-26 19:01:46 |