OVERVIEW: These techniques extrapolate past data into the future. The premise is that there is some underlying pattern in the value of the variable being forecast.
GENERAL APPROACH: To choose the appropriate technique, first do the following:
STEP 1: Date Examination - Graphical Analysis
What is the pattern in the data?
Try plotting the historical data on a chart
Look for a trend line by fitting a simple line or curvilinear trend curve to the data in a freehand method.
Look for cyclical patterns, if any, and/or major turning points
Choose an appropriate technique. Table A2 presents different techniques and their capabilities in recognizing various patterns. For example,
If the trend line is horizontal or reasonably stable (little or no slope), try simple versions of:
Moving averages, or
If slope is evident, try
Double moving averages, or
Double exponential smoothing
Fluctuating patterns can be handled with:
Decompositional models, or more complex approaches such as: