Novel Framework for Dengue Classification and Early Recovery using Machine Learning Algorithms

Dengue fever is a fast-emerging public health hazard that primarily affects tropical and sub-tropical regions. Arthropods that feed on blood, such as fleas, parasites, and mosquitoes, are the carriers of this disease, which can infect humans and cause significant illness. Five serotypes of the dengue virus—DEN V1, DEN V2, DEN V3, DEN V4, and DEN V5 are spread by Aedes aegypti mosquitoes. Currently, this illness is spreading to many other countries. A delayed diagnosis combined with the wrong method of treatment increases the chance of death. Based on the severity of the infection, dengue is classified into Dengue Hemorrhagic Fever (DHF) or Dengue Shock Syndrome (DSS) which are categorized into febrile, critical, and recovery phases. A framework model is essential to analyze the severity of the infection and identify the phase of severity. Among various machine learning algorithms, this study uses Support Vector Machine and Random Forest for their proficiency in handling complex data. Through rigorous feature engineering and data preprocessing, critical information for precise discrimination was extracted and examined for DHF phases. Experimental results demonstrate both models are competent, but SVM is more suitable for the proposed prediction model with an accuracy 95%, precision of 93%, recall of 94%, and F-measure metrics of 94%.