Location
Online
Event Website
https://hicss.hawaii.edu/
Start Date
3-1-2023 12:00 AM
End Date
7-1-2023 12:00 AM
Description
Arboviral diseases are common worldwide. Infection with arboviruses can lead to serious health problems, even death in severe cases. Such health problems can be prevented by the early and correct detection of these arboviruses, but this is challenging due to the overlap of their symptoms. In this work, we benchmark different Machine Learning (ML) models to classify two types of arboviruses. We propose two distinct binary models: (i) a model to classify if the patient has arbovirus or another disease; and (ii) a model to classify if the patient has Dengue or Chikungunya. We configure and evaluate several ML models using hyperparameter optimization and feature selection techniques. The Random Forest and XGboost tree-based models present the best results with over 80% recall in the Chikungunya and Inconclusive classes.
Recommended Citation
Da Silva Neto, Sebastião Rogerio; Tabosa, Thomás; Medeiros Neto, Leonides; Teixeira, Igor Vitor; Sadok, Sara; De Souza Sampaio, Vanderson; and Endo, Patricia Takako, "Binary Models for Arboviruses Classification Using Machine Learning: A Benchmarking Evaluation" (2023). Hawaii International Conference on System Sciences 2023 (HICSS-56). 2.
https://aisel.aisnet.org/hicss-56/hc/process/2
Binary Models for Arboviruses Classification Using Machine Learning: A Benchmarking Evaluation
Online
Arboviral diseases are common worldwide. Infection with arboviruses can lead to serious health problems, even death in severe cases. Such health problems can be prevented by the early and correct detection of these arboviruses, but this is challenging due to the overlap of their symptoms. In this work, we benchmark different Machine Learning (ML) models to classify two types of arboviruses. We propose two distinct binary models: (i) a model to classify if the patient has arbovirus or another disease; and (ii) a model to classify if the patient has Dengue or Chikungunya. We configure and evaluate several ML models using hyperparameter optimization and feature selection techniques. The Random Forest and XGboost tree-based models present the best results with over 80% recall in the Chikungunya and Inconclusive classes.
https://aisel.aisnet.org/hicss-56/hc/process/2