Application of Holt exponential smoothing and ARIMA method for data population in West Java

The Indonesian Operations Research Association (IORA) - International Conference on Operations Research 2016

Supriatna, A.^1 ; Susanti, D.^1 ; Hertini, E.^1

Department of Mathematics, Faculty of Mathematics and Natural Science, Padjadjaran University, West-Java, Indonesia^1

关键词: Data patterns;    Data population;    Exponential smoothing;    West javas;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/166/1/012034/pdf
DOI  :  10.1088/1757-899X/166/1/012034

来源: IOP

【 摘 要 】

One method of time series that is often used to predict data that contains trend is Holt. Holt method using different parameters used in the original data which aims to smooth the trend value. In addition to Holt, ARIMA method can be used on a wide variety of data including data pattern containing a pattern trend. Data actual of population from 1998-2015 contains the trends so can be solved by Holt and ARIMA method to obtain the prediction value of some periods. The best method is measured by looking at the smallest MAPE and MAE error. The result using Holt method is 47.205.749 populations in 2016, 47.535.324 populations in 2017, and 48.041.672 populations in 2018, with MAPE error is 0,469744 and MAE error is 189.731. While the result using ARIMA method is 46.964.682 populations in 2016, 47.342.189 in 2017, and 47.899.696 in 2018, with MAPE error is 0,4380 and MAE is 176.626.