EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Leveraging a Random Forest Classifier and SVMSMOTE for an Early-stage Dengue Prediction

Yulianti Paula Bria, Paskalis Andrianus Nani, Yovinia Carmeneja Hoar Siki, Natalia Magdalena Rafu Mamulak, Emiliana Metan Meolbatak, Robertus Dole Guntur

Engineering, Technology & Applied Science Research · 04.06.2025 · 15 (3) · 23436-23442

DOI: 10.48084/etasr.10762 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Engineering, Technology & Applied Science Research, online erschienen am 04.06.2025. Verfasst von Yulianti Paula Bria und 5 weiteren Autor:innen. Die Arbeit wurde bisher 5-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Dengue is a significant global health issue, and in its severe form, it can be fatal. Accurate early-stage prediction tools are crucial in resource-limited areas to prevent dengue's progression to a severe state. This study aims to develop machine learning classifiers for dengue to assist medical personnel in differentiating the latter from other diseases, thereby helping with its earlier prognosis. Early-stage dengue classifiers were developed using medical records collected from two hospitals in East Nusa Tenggara Province, Indonesia. Eight machine learning techniques were leveraged to develop the classifiers, including Logistic Regression (LR), Linear Discriminant Analysis (LDA), Support Vector Machine (SVM), K-nearest Neighbors (KNN), Naïve Bayes (NB), Classification and Regression Tree (CART), Random Forest (RF), and extreme Gradient Boosting (XGBoost). To address the imbalance in the dataset, we utilized the SVM Synthetic Minority Oversampling Technique (SVMSMOTE) was utilized. The dataset was finalized through the expertise of 15 medical doctors and insights gathered from four Indonesian digital health platforms. The key findings of this study are: i) important features for early-stage dengue prediction include fever, duration of fever, headache, arthralgia, myalgia, nausea, shivering, loss of appetite, bitter mouth, temperature, and age, ii) machine learning techniques, including RF, NB, KNN, and XGBoost, were found to be suitable for dengue prediction, and iii) RF combined with SVMSMOTE, outperformed other techniques, achieving an accuracy of 94.99% and an F1-score of 85.65% for early-stage dengue prediction.

Zitieren

APA

Bria, Y. P., Nani, P. A., Siki, Y. C. H., Mamulak, N. M. R., Meolbatak, E. M., & Guntur, R. D. (2025). Leveraging a Random Forest Classifier and SVMSMOTE for an Early-stage Dengue Prediction. Engineering, Technology & Applied Science Research, 15(3), 23436-23442. https://doi.org/10.48084/etasr.10762

Vancouver

Bria YP, Nani PA, Siki YCH, Mamulak NMR, Meolbatak EM, Guntur RD. Leveraging a Random Forest Classifier and SVMSMOTE for an Early-stage Dengue Prediction. Engineering, Technology & Applied Science Research. 2025;15(3):23436-23442. doi:10.48084/etasr.10762

BibTeX
@article{Bria2025,
  author = {Bria, Yulianti Paula and Nani, Paskalis Andrianus and Siki, Yovinia Carmeneja Hoar and Mamulak, Natalia Magdalena Rafu and Meolbatak, Emiliana Metan and Guntur, Robertus Dole},
  title = {Leveraging a Random Forest Classifier and SVMSMOTE for an Early-stage Dengue Prediction},
  journal = {Engineering, Technology & Applied Science Research},
  year = {2025},
  volume = {15},
  number = {3},
  pages = {23436--23442},
  doi = {10.48084/etasr.10762}
}
Kostenlos starten →