Using the ets function

While moving averages re extremely useful, they are only one component of what is known as an exponential smoothed state space model, which has many options to define the optimal smoothing factor, as well as enabling you to define the type of trend and seasonality via the parameters. To implement this model we will use the ets() function from the forecast package to model the Not-Covered Percent variable for the "ALL AGES" category.

The ets() function is flexible in that it can also incorporate trend, as well as seasonality for its forecasts.

We will just be illustrating a simple exponentially smoothed model (ANN). However, for completeness, you should know that you specify three letters when calling the ets() function, and you should be aware of what each letter represents. Otherwise, it will model based upon the default parameters.

Here is the description as specified by the package author, Hydman:

• The first letter denotes the error type ("A", "M", or "Z")

• The second...