VOLUME 19 (Supplement)

Using machine learning models for near-term forecasting of dengue cases in Quezon City, Philippines

Dengue is a vector-borne illness that has been a long-standing health concern in the Philippines. The disease is endemic in all regions of the country, especially in highly urbanized areas like the National Capital Region (NCR), and particularly in Quezon City (QC). The local government of QC, through the Quezon City Epidemiology and Surveillance Division (QCESD), uses a real-time monitoring system that tracks epidemic levels, reports cases by barangay, and identifies high-risk areas requiring public health interventions. The QCESD continues to enhance its system by finding ways to generate actionable insights from the epidemiological data collected. This study aims to contribute to this effort by supplementing surveillance with dengue forecasting tools. Specifically, machine learning models capable of predicting dengue incidence using epidemiological, meteorological, environmental, and socioeconomic data were implemented. Different machine learning models were explored and their performance assessed in terms of predictive accuracy and robustness. Aside from case forecasts, the factors with the greatest influence on predicted case trajectories were also identified.