Aplikasi Autoregressive Integrated Moving Average (ARIMA) untuk Meramalkan Jumlah Demam Berdarah Dengue (DBD) di Puskesmas Mulyorejo

ARIMA was one of a forecasting method of time series if independent variable be ignored, it would use the past and present value as a dependent variable. The accuracy of ARIMA forecasting method was good to produce short-term forecasting. The advantages of ARIMA method than other method was this method didn’t require the data pattern so it could be used for all kinds of data pattern, so it could be applied in cases of dengue hemorrhagic fever (DHF) in Mulyorejo Public Health Center. This study was to determine the best forecasting model as well as to predict and analyze the results of forecasting number of dengue hemorrhagic fever in Mulyorejo Public Health Center. The data was monthly number of dengue hemorrhagic fever patients in Mulyorejo Public Health Center from January 2010 to February 2016 (a total of 74 plots data). The results were the number of dengue hemorrhagic fever cases in Mulyorejo Public Health Center could be predicted with ARIMA model (1,0,0), thought based on diagnostics test the ARIMA model met all tests but the forecasting number of dengue hemorrhagic fever cases in years 2016–2017 showed a downward trend, and in 2017 was fl at, while MAPE and MAE amounted to 63.026% and 1.89%, the value of the error was large enough which indicated that less accurate forecasting. DHF data had a lot of missing data caused big value of MAPE and MAE so must be transformed by series mean method. DHF data was trend and seasonal so winters exponential smoothing with ordinary least square was better than ARIMA to get small error.