Forecasting hotspots in East Kutai, Kutai Kartanegara, and West Kutai as early warning information

1st International Conference on Tropical Studies and Its Application

Wahyuningsih, S.^1 ; Goejantoro, R.^1 ; Rizki, N.A.^1

Study Program of Statistics, Department of Mathematics, Faculty of Mathematics and Natural Sciences, Mulawarman University, Samarinda, Indonesia^1

关键词: Additive decomposition;    Autoregressive moving average;    Decomposition methods;    Decomposition model;    Exponential smoothing;    Multiplicative decomposition;    Smoothing techniques;    Times series models;   

Others  :  https://iopscience.iop.org/article/10.1088/1755-1315/144/1/012022/pdf
DOI  :  10.1088/1755-1315/144/1/012022

来源: IOP

【 摘 要 】

The aims of this research are to model hotspots and forecast hotspot 2017 in East Kutai, Kutai Kartanegara and West Kutai. The methods which used in this research were Holt exponential smoothing, Holt's additive dump trend method, Holt-Winters' additive method, additive decomposition method, multiplicative decomposition method, Loess decomposition method and Box-Jenkins method. For smoothing techniques, additive decomposition is better than Holt's exponential smoothing. The hotspots model using Box-Jenkins method were Autoregressive Moving Average ARIMA(1,1,0), ARIMA(0,2,1), and ARIMA(0,1,0). Comparing the results from all methods which were used in this research, and based on Root of Mean Squared Error (RMSE), show that Loess decomposition method is the best times series model, because it has the least RMSE. Thus the Loess decomposition model used to forecast the number of hotspot. The forecasting result indicatethat hotspots pattern tend to increase at the end of 2017 in Kutai Kartanegara and West Kutai, but stationary in East Kutai.