![]() Notice how the prediction intervals widen with subsequent forecasts. The forecasts (green diamonds) increase at a rate equal to the final trend estimate. I’ve allowed the software to estimate the level (0.599) and trend (0.131) smoothing constants from the data to optimize the fit. As you can see in the chart, the time series data have a trend. In the example below, we’re using double exponential smoothing to model monthly computer sales. A popular extension for this method adds a dampening component to the forecasts, causing the forecasts to level out over time to avoid overly optimistic long-term forecasts. Higher values place more weight on recent observations, allowing the trend component to react more quickly to changes in the trend.įorecasts for this method change at a constant rate equal to the final value of the trend component. Double exponential smoothing is also known as Holt’s Method.Īs with alpha, beta can be between 0 and 1, inclusive. To model trends, DES includes an additional parameter, beta (β*). This method models dynamic gradients because it updates the trend component for each observation. Double Exponential Smoothing (DES)ĭouble exponential smoothing can model trend components and level components for univariate times series data. The prediction intervals indicate the uncertainty surrounding the predictions.Īs we move on to double and triple exponential smoothing, notice how each method adds components to the model, extending its functionality. The forecasts (green diamonds) are all constant values at the final estimate of the level component. Unsurprisingly, the accuracy measures are even lower (better) for this model than for the previous two models. The software estimates that the optimal alpha smoothing constant is 0.834. Now, let’s allow the software to find the optimized alpha parameter and generate forecasts. I’ll define these measures in a later post, but lower values represent a better fitting model.īased on the accuracy measures, the model using an alpha of 0.8 provides a better fit. ![]() Similarly, forecasts from this model place more weight on recent observations than the other model. This line adjusts to the changing conditions more rapidly. The larger alpha (smaller the damping factor), the closer the smoothed values are to the actual data points.Notice how the time series plot using 0.8 has a more jagged fitted line (red) than the other graph. Repeat steps 2 to 8 for alpha = 0.3 and alpha = 0.8.Ĭonclusion: The smaller alpha (larger the damping factor), the more the peaks and valleys are smoothed out. The smoothed value for the second data point equals the previous data point.ĩ. ![]() Excel cannot calculate the smoothed value for the first data point because there is no previous data point. As a result, peaks and valleys are smoothed out. Click in the Output Range box and select cell B3.Įxplanation: because we set alpha to 0.1, the previous data point is given a relatively small weight while the previous smoothed value is given a large weight (i.e. The value (1- α) is called the damping factor.Ħ. Literature often talks about the smoothing constant α (alpha). Click in the Damping factor box and type 0.9. Click in the Input Range box and select the range B2:M2.ĥ. Select Exponential Smoothing and click OK.Ĥ. Note: can't find the Data Analysis button? Click here to load the Analysis ToolPak add-in.ģ. On the Data tab, in the Analysis group, click Data Analysis.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |