Perceptron-based dengue risk mapping in Cagayan de Oro City

Abstract

The rising incidence of dengue cases in the Philippines continues to pose a public health concern. Risk mapping through machine learning algorithms has been effective in surveilling areas vulnerable to dengue outbreaks. In this paper, we compared two risk formulas using weights obtained from a Perceptron model. These weight predictions were used in mapping dengue risk on a granular scale. The resulting risk map of Cagayan de Oro City in 2022 reveals high-risk zones concentrated in barangays near the urban core. Built-up areas were the primary contributors to heightened risk. Furthermore, the Perceptron-based risk map resulted in high accuracy, outperforming pairwise comparisons to quantify risk factors. This framework can be used to implement targeted dengue control strategies.