Public administration data is sometimes extrapolated through exponential smoothing. Sometimes such data may undergo a level shift because of a policy decision. The slope of the curve formed by connecting the periodic observations increases or decreases significantly for a brief period, thereafter returning to a slope similar to the slope preceding the policy change. This discontinuity might be called a ramp or a step. Forecasts made with exponential smoothing immediately before, during, or immediately after the ramp or step may be considerably inaccurate unless adjusted. A technique called adjusted exponential smoothing is proposed to reduce or eliminate the inaccuracy of forecasts made under such circumstances when the ramp or step arises from a planned policy decision. An empirical study is conducted to determine whether the proposed technique constitutes an improvement over other exponential smoothing techniques. The empirical study shows that the proposed technique improves the accuracy of forecasts when planned level shifts subsequently actually occur. Guidelines are provided for using the technique.
| Identifer | oai:union.ndltd.org:vcu.edu/oai:scholarscompass.vcu.edu:etd-6169 |
| Date | 01 January 1994 |
| Creators | Williams, Daniel Wayne |
| Publisher | VCU Scholars Compass |
| Source Sets | Virginia Commonwealth University |
| Detected Language | English |
| Type | text |
| Format | application/pdf |
| Source | Theses and Dissertations |
| Rights | © The Author |
Page generated in 0.0019 seconds