Dengue Prediction in Bangladesh Using Machine Learning

Abstract

Dengue fever continues to be a health issue, in Bangladesh causing harm and even deaths during frequent outbreaks. This study employs machine learning techniques to forecast dengue outbreaks by analyzing data from 820 cases across hospitals in Bangladesh. By using algorithms like Logistic Regression, Random Forest, Gradient Boosting, Support Vector Machine, K Nearest Neighbors and XGBoost the study aims to determine the method for prediction. The findings reveal that logistic regression outperformed algorithms in terms of accuracy. To improve the models performance various data preprocessing techniques were employed such as LabelEncoder for encoding labels ADASYN oversampling technique for handling data filling values with median and mode values StandardScaler for scaling data and outlier capping to handle extreme values. The insights gained from this research could play a role in enhancing public health strategies aimed at controlling and preventing dengue fever, in Bangladesh.