Check out our Simple Exponential Smoothing tutorial below. The playlist contains tutorials that teach you how to utilize the optimization switch, calibrate with a training set, and calculate an in-sample forecast using NumXL.
In this video, we show you how to use Brown's simple exponential smoothing function in NumXL with an optimization switch for smoothing factors.
In this video, we show you how to designate a portion of data to calibrate your smoothing factor for Simple Exponential Smoothing.
Hello and welcome to the exponential smoothing tutorial series. In our last tutorial we showed you how you can enable the built-in optimizer in the simple exponential smoothing function to find the optimal values for the smoothing parameters.
In this tutorial we will demonstrate how to designate a portion of data for calibrating the smoothing parameter, then we will use the found value with the rest of the sample data. \
For the sample data we will continue using the two years sales data from our hypothetical company, and we'll use the first year to calibrate the simple exponential smoothing model.
First let us disable the optimization for the smooth time series. In our other videos we reference cell D2 as the optimization switch, so select cell D2 and type in false or 0.
Select the cells in column D that correspond to the training set and delete the formula.
Now we are ready for calibrating the smoothing parameter. Select the cell in D1 and type in =SESMTH(.
Once you find the function click on the FX button found on the left side of the equation toolbar. This will invoke the function arguments dialog box for the simple exponential smoothing function.
For input data select the cell range in the training set. In this case it is C9 to C20. Set the chronical order of the time series to true or 1, this will signal that the first observation in the input data corresponds to the earliest date.
You can leave alpha blank or enter any value between 0 and 1 this number will be used as a starting value for the optimizer.
Turn on the optimizer by typing in true or 1, set the forecast time to 1, so that the whole training set is used. In return type, enter 1 then click OK.
The optimal value of the smoothing factor in the training set is returned to D1. We have the automatic calculation turned on so the smooth time series in D21 to D36 are updated automatically. Here you can note the change in the forecast performance measures.
That's all for now, thank you for watching!
In this video, we show you how to use Brown's simple exponential smoothing function in NumXL without the optimization switch for smoothing factors.
In this video, we show you how to calculate a fitted or in-sample forecast for your Simple Exponential Smoothing function in NumXL.