Abstract:
The exponential smoothing forecasting method is one of the time series data forecasting
methods. This study uses non-oil and gas export data in Indonesia for the period January
2015-May 2021 indicating that there are elements of trends and seasonality. Based on the
characteristics of the time series, the data obtained were then analyzed using Holt's double
exponential smoothing method and triple exponential smoothing consisting of
multiplicative and Holt-Winters additives. The best model selection can be done by
choosing the minimum forecast error value. This study shows that the additive Holt-
Winters method is the best forecasting method that has a minimum forecast error value.
The forecast results using this method indicate that the value of non-oil exports will
increase for the next period.
