Time Series Analysis and Forecasting on the Dengue Case Incidence in Iloilo using SARIMA Models

Introduction

Dengue remains a major public health concern in the Philippines, with seasonal surges during the rainy season placing significant strain on public healthcare systems. This study models weekly dengue incidence in Iloilo, a province in Western Visayas, using time series analysis to identify short-term and multi-year seasonal patterns. After data pre-processing and statistical testing, several SARIMA models were fitted and evaluated using the Akaike Information Criterion (AIC). The best-fit model, SARIMA(0, 1, 1) × (0, 1, 1)_156, suggests that weekly dengue incidence is influenced by recent case counts and three-year seasonal patterns. Forecasts from this model project surges around mid-2023 and 2024, aligning with known annual patterns despite the triannual seasonal specification. These results may inform the timing of seek-and-destroy operations on mosquito breeding sites and hospital resource allocation. However, wide forecast confidence intervals also highlight the limitations of relying solely on historical statistical models. Combining time series models with real-time surveillance and expert recommendations remains essential for effective and timely public health response.

Files

• PDF: Research paper
• R File: Codes for the time series analysis and forecasting
• disease_pidsr_totals.csv and location.csv: Datasets used in the study

Credits

This project was created by Sted Cheng, Ivan De Leon, and Annika Montemayor and submitted as a requirement for the course MATH 271.2: Advanced Time Series and Forecasting taken in the intersession (midyear) term of AY 2025-2026 in Ateneo de Manila University.