Forecasting Diarrhea Incidence Using ARIMA, SARIMA, and Autotuned SARIMA Models

Dedy Hidayat Kusuma (Departement of Bussines and Informatics, Politeknik Negeri Banyuwangi, Banyuwangi, Indonesia)
Moh Nur Shodiq (Departement of Bussines and Informatics, Politeknik Negeri Banyuwangi, Banyuwangi, Indonesia)
Herman Yuliandoko (Departement of Bussines and Informatics, Politeknik Negeri Banyuwangi, Banyuwangi, Indonesia)
Muh. Fuad Al Haris (Departement of Bussines and Informatics, Politeknik Negeri Banyuwangi, Banyuwangi, Indonesia)
Sri Andayani (Puskesmas Sumberberas, Dinas Kesehatan Kabupaten Banyuwangi, Banyuwangi, Indonesia)

Jurnal Krisnadana
Vol 5 No 1 (2025): Jurnal Krisnadana- in Progress September-October 2025

Diarrhea remains a major public health concern in developing countries, including Indonesia, contributing significantly to morbidity and mortality among children under five years old. Its incidence fluctuates due to environmental factors, making accurate forecasting essential for effective health resource planning. This study aims to predict monthly diarrhea cases at Community Health Center Sumberberas, Banyuwangi, from January 2023 to May 2025 using three time series models: ARIMA, SARIMA, and SARIMA with autotuning. A quantitative approach was applied, consisting of data preprocessing, model construction, evaluation, and comparison. Model performance was assessed using Mean Absolute Error (MAE), Root Mean Squared Error (RMSE), Akaike Information Criterion (AIC), and Bayesian Information Criterion (BIC). The results show that the SARIMA with autotuning model achieved the lowest MAE, RMSE, AIC, and BIC values, demonstrating superior accuracy and model fit compared to ARIMA and manually tuned SARIMA. These findings indicate that SARIMA with autotuning provides the most reliable forecasts for short-term diarrhea incidence, supporting data-driven decision-making and the development of early warning systems for disease prevention and control.