We will now explore the methods of building the forecasting model using the Holt-Winters method with the HoltWinters
function, which belongs to the forecasting
package. This function computes the Holt-Winters filtering for a given time series data, and the unknown parameters are determined by minimizing the squared prediction error.
Apart from passing the time series data as an input, the other important parameters that need to be passed to this function are the alpha, beta, and gamma, as follows:
alpha
: The parameter of the Holt-Winters filtersbeta
: This is used for the trend component; when set to false, the function will do exponential smootheninggamma
: This is used for the seasonal component; when set to false, the nonseasonal component is fitted
Let's execute the following code where the trend component is set to FALSE
, and hence, exponential smoothening will be performed on the dataset. We have set the gamma
value as 0.5
:
h_model=HoltWinters(ts, beta...