Predicting Case Fatality of Dengue Epidemic: Statistical Machine Learning Towards a Virtual Doctor

Dengue fever is a self-limiting communicable viral disease, transmitted through mosquito bites. Its Case Fatality Grade (CFG) varies across population due to variations in load, immunity of the patient, early diagnosis, and availability high-end treatment facility. This study describes an initial effort automate process CFG predictions. Two established Statistical Machine Learning (SML) algorithms, Multiple Linear Regressions (MLR) Multinomial Logistic (MnLR), are combined substitute existing Deep methods for clinical decision making. We consider vector eleven sign-symptoms (independent variables), each weighted between [0,1] on 3-point scale - ‘Mild’ (CFG<=0.33), ‘Moderate’ (0.33<CFG< 0.66), ‘Severe’ (CFG>0.66). Results show that both classifiers effective screening with similar accuracy levels (68% MLR versus 72% MnLR) although precision far superior MnLR (88%) than (61%). futuristic step towards (ML) aided diagnostic paradigms, as alternative computationally intensive Artificial Intelligence.