The accuracy of forecasting results of the Box-Jenkins method for time series analysis on the number of pneumonia patients

Abstract

This study was quantitative with a non-reactive or unobstructed approach using Time Series analysis with the ARIMA Box-Jenkins method on secondary data. From secondary data, 1, 3, and 6 monthly data simulations were carried out. Each simulation data was divided into two groups: the first group of initialization data for 2016 – 2019 and the second group of actual data for 2020. The aim of this study was to determine the accuracy of forecasting results using the Box- Jenkins method on the number of pneumonia sufferers at Kamonji Public Health Center, Palu. The 1-monthly data simulation amounted to 48-time series as initialization data, obtained the appropriate forecasting model, namely ARIMA (1,1,1), then forecasting was done, and the results obtained were 289,166 patients with pneumonia. There are 16-time series simulations for 3-monthly data as initialization data and 6-monthly simulations totaling two-time series for initialization data, not finding a suitable model for forecasting. In conclusion, no data simulation gets the right result on the number of pneumonia sufferers because it gets more forecasting results than the actual data. Suggestions that can be given in this research are to use data that is more than five years old.