DenGuess: Design and Development of a Machine Learning Web-Based Application for Dengue Outbreak Prediction in City of Koronadal Using Random Forest Model Classifier

Abstract

Dengue outbreaks remain a major public health problem in Koronadal City with limited technology for monitoring the disease. Health workers are often challenged in accessing and preventing outbreaks due to the lack of available tools. While Random Forest models have been shown to be useful for predicting dengue, their practical application remains still limited. In this study, the researchers developed a machine learning web application to predict dengue outbreaks for five (5) selected barangays in Koronadal weekly, using the epidemiological and meteorological data from 2020 to 2025. The model’s performance was measured using a) real-time simulation accuracy, b) precision, c) recall, and d) F1-scores, and the website was evaluated using the user interface and user experience. The findings demonstrate that the model provides balanced and satisfactory predictions, and the website scores high on all criteria, confirming the usability and reliability of the website. This system provides timely forecast information to support the decision-making process, create more awareness among the community and increase preparedness in the city. Researchers suggest further systematizing the UI and UX of the website and to integrate the system into the city’s public health infrastructure to improve data collection, validation, and engagement within the community.