From the course: Business Analytics: Forecasting with Trended Baseline Smoothing
Unlock the full course today
Join today to access over 22,600 courses taught by industry experts or purchase this course individually.
Initialize the trend forecasts
From the course: Business Analytics: Forecasting with Trended Baseline Smoothing
Initialize the trend forecasts
- [Instructor] To this point in the course I have not shown in any detail part of building a forecast on exponential smoothing, and that's initialization. Recall these two points. One, you have to supply the first forecast because exponential smoothing depends on a prior observation to generate each new forecast. But there is no observation that proceeds time one. Else it would not be time one. So there was no naturally occurring time zero on which we can base a forecast for time one. Point two, an easy and convenient way around this problem, is to assign the value of the first observation to be the first forecast. With both an observation and a forecast available for time two, the smoothing process can start in earnest at time two. Let's explore this method a bit more along with a couple more ways to initialize a forecast. I'm in exercise file 02_04 in the initialize level simple worksheet. Here there is no forecast for time one in cell three, because no time zero observation exists…
Practice while you learn with exercise files
Download the files the instructor uses to teach the course. Follow along and learn by watching, listening and practicing.
Contents
-
-
-
-
(Locked)
Distinguish between a level component and a trend component3m 28s
-
(Locked)
The trend constant compared to the level constant3m 25s
-
(Locked)
Compare smoothing and error correction forms6m 30s
-
(Locked)
Initialize the trend forecasts5m 2s
-
(Locked)
Build the full worksheet and optimize with Solver6m 50s
-
(Locked)
-
-