Comparison of machine learning algorithms for dengue virus (DENV) classification

Exebio-Chepe Y.V.; Bravo-Ruiz J.A.; Tuesta-Monteza V.A. · Journal of Applied Research and Technology (2024) · DOI: 10.22201/icat.24486736e.2024.22.5.2422

Resumen

Dengue transmitted by the Aedes aegypti mosquito, requires accurate classification of cases for effective management, which is currently a gap of study with a particular case in Peru. The research focuses on leveraging machine learning algorithms to improve diagnosis and streamline control strategies concerning dengue transmission cases. Using a dataset from a public hospital, covering 21,157 cases classified by period, outcome, sex, age, symptoms, and origin (autochthonous or imported), the study performed a comparative analysis of support vector machine, random forest, and artificial neural network algorithms. The data set was divided into 70% (14,809 cases) for training and 30% (6,348 cases) for testing. The results revealed that artificial neural network came out on top with 86.47% accuracy and 92.91% recall in classifying dengue-related cases. It is concluded that the implementation of support vector machine proved to be sensitive of 99.05%, highlights the effectiveness in dengue case classification.

Instituciones: Universidad Señor de Sipán