A machine learning approach: predicting the number of dengue patients in Bangladesh based on climate changing data

Abstract

Aedes aegypti mosquitoes are the vector of dengue fever, which can cause fatalities. Symptoms range from mild flu-like symptoms to severe diseases like shock syndrome and dengue hemorrhagic fever. In light of climate-related data, this work presents a reliable machine learning model for estimating the number of Dengue patients in Bangladesh. Various models, such as regression models and KNN, Decision Tree, and Random Forest for classification, are used in supervised learning with labeled data. The impact of dengue in 2022–2023 emphasizes how urgent it is to address this health catastrophe, as an increasing number of cases and deaths are being caused by collective neglect. In spite of previous studies, this analysis provides insightful information based on the most recent data relevant to Bangladesh. Interestingly, Random Forest performed quite well in both regression and classification, whereas Decision Tree was very effective in the former, showing an excellent f1 score of 0.85 and a marginally lower accuracy of 84.79% in comparison to Random Forest's 86.20%.