Machine learning based prediction of dengue risk zones in Bangladesh based on weather data

Abstract

This work proposes a unique method for identifying dengue risk zones in Bangladesh using machine learning algorithms and large weather data. Dengue fever, a mosquito-borne illness that is common in tropical areas, has a complicated interplay with climate factors. Using GNB, Random Forest Classifier, Decision Tree Classifier, and Voting Classifier, Our Machine learning model uses Gradient Boosting Classifier and Logistic Regression to uncover subtle patterns in rain, temperature, and oxygen. The study combines past weather data with reported dengue cases, employing a number of machine learning methods to determine connections between environmental variables and illness incidence. Our algorithm delivers nuanced risk evaluations by applying a complex ensemble of classifiers, classifying regions as "High Risk," "No Risk," "Low Risk," and "Moderate Risk." It allows for focused public health interventions, more effective use of resources, and proactive dengue epidemic control. The proposed machine learning-based prediction model not only tackles the current threat of dengue in Bangladesh, but it also serves as a tough tool that can be adapted to changing climatic dynamics. This study adds to the larger conversation on the connection of data science and public health by providing an adjustable and dynamic framework for minimising the effect of vector-borne illnesses in climate-vulnerable areas.