Specifying Time Intervals

Statistical forecasting requires that the historical input data be organized as time seriesan aggregation of transactional data into specified time intervals and sorted according to unique combinations of the default attributes (BY variables), where successive measurements are taken at points equally spaced in time. SAS Visual Forecasting satisfies this requirement in three ways:

• The input data is already in time series form, and is used without any alteration in the time dimension.
• The input data is already in time series form, but is aggregated over time to a different interval (lower frequency), which you specify.
• The input data is transactional (recorded at no particular interval) and is aggregated over time to form time series at an interval which you specify.

Define the time interval specifications when you first create your project using the Assign Variables window. For more information, see Assigning Variable Roles.

SAS Visual Forecasting represents time intervals using a very flexible notation based on a set of basic intervals and optional settings. These settings include the multiplier (affects the length of the interval) and the shift (affects the starting point of the interval). For most time series, the start, length, and end of every observation is identifiable once the time interval is specified in this way. See Time Variable for a description of each setting.

For the time variable, SAS Visual Forecasting detects the best default settings for the time interval, multiplier, and shift. They are determined by analysis of the time variable values. For input data that are already in time series form, you can often proceed with these default values. You can change them if you know the interval and disagree with the defaults. You can choose any time interval specification that is of lower frequency than the default setting. For transactional datatimestamped data collected over time at no particular frequency. Some examples of transactional data are point-of-sale data, inventory data, call center data, and trading data., choose an interval specification in which most or all time periods encompass times at which transactions were recorded.

The examples in the following table show how the values that you specify for the interval, multiplier, and shift work together. For the interval name using SAS code, the m.s suffix indicates the multiplier and shift settings for that time interval. In the examples shown, YEAR2.7 indicates a time interval set to YEAR, multiplier set to 2, and shift set to 7.

Time Interval, Multiplier, and Shift Examples

Interval Name (in SAS code format)	Default Starting Point	Shift Period	Example
YEARm.s	January 1	Months	YEAR2.7 specifies an interval of every two years with the first month of the year starting in July.
SEMIYEARm.s	January 1 July 1	Months	SEMIYEAR.3 specifies six-month intervals spanning March-August and September-February.
QTRm.s	January 1 April 1 July 1 October 1	Months	QTR.3 specifies three-month intervals starting on March 1, June 1, September 1, and December 1.
SEMIMONTHm.s	First and 16th of each month	Semimonthly periods	SEMIMONTH2.2 specifies intervals from the 16th of one month through the 15th of the next month.
MONTHm.s	First of each month	Months	MONTH2.2 specifies February-March, April-May, June-July, August-September, October-November, and December-January of the following year.
TENDAYm.s	First, 11th, and 21st of each month	Ten-day periods	TENDAY4.2 specifies four 10-day periods starting at the second 10-day period.
WEEKm.s	Each Sunday	Days (1=Sunday . . . 7=Saturday)	WEEK6.3 specifies six-week intervals starting on Tuesdays.
DAYm.s	Each day	Days	DAY3 specifies three-day intervals starting on Sunday.
HOURm.s	Start of the day (midnight)	Hours	HOUR8.7 specifies eight-hour intervals starting at 6:00 a.m., 2:00 p.m., and 10:00 p.m.