TEKNIK PERAMALAN MENGGUNAKAN METODE PEMULUSAN EKSPONENSIAL HOLT-WINTERS

Abstract

Time series forecasting is a method used to determine what might happen in the future based on information obtained in the past. One method used in time series forecasting is a Holt-Winters exponential smoothing method. This method can be used for time series data with trend and seasonality components. This method is based on three smoothing equations: overall smoothing, trend, and seasonal components. Holt-Winters exponential smoothing method consists of multiplicative and additive seasonality models. The method of this research is literature study by collecting and studying references that are relevant to the idea of this research, and then applying the Holt-Winters exponential smoothing method into data. The results of this research show that the multiplicative seasonality model of Holt-Winters exponential smoothing method can be used if data represent an increase in long-term and seasonal fluctuations which is the increasingly bigger with the increasing of observation time periods. These patterns identify the non-stationary of mean and variance. While, the additive seasonality model can be used if data show an increase in long-term and seasonal fluctuations that are relatively constant with the increasing of observation time. Time series forecasting is a method used to determine what might happen in the future based on information obtained in the past. One method used in time series forecasting is a Holt-Winters exponential smoothing method. This method can be used for time series data with trend and seasonality components. This method is based on three smoothing equations: overall smoothing, trend, and seasonal components. Holt-Winters exponential smoothing method consists of multiplicative and additive seasonality models. The method of this research is literature study by collecting and studying references that are relevant to the idea of this research, and then applying the Holt-Winters exponential smoothing method into data. The results of this research show that the multiplicative seasonality model of Holt-Winters exponential smoothing method can be used if data represent an increase in long-term and seasonal fluctuations which is the increasingly bigger with the increasing of observation time periods. These patterns identify the non-stationary of mean and variance. While, the additive seasonality model can be used if data show an increase in long-term and seasonal fluctuations that are relatively constant with the increasing of observation time