@article{article_1813790, title={Improved arbovirus suspected case analysis via ensemble methods with parameter tuning: Insights from SISA dataset}, journal={Pamukkale Üniversitesi Mühendislik Bilimleri Dergisi}, year={2025}, DOI={10.65206/pajes.24040}, author={Doğan, Alican}, keywords={Makine öğrenimi, Arbovirüs Enfeksiyonu, Rastgele Orman, Hastane yatış durumu}, abstract={Hospital admission necessity of a patient who is under care for the possibility of arbovirus infection is a critical decision for healthcare practitioners. Medical staff may experience stress when making this decision due to the potential risks it poses to the broader community. Current capacities for diagnosis can be confusing. For this reason, data mining approaches have been proven to be highly effective in the diagnosis of diseases as well as in many other fields. As many research studies suggest, they can also be used to decide whether a patient with arbovirus infection should be hospitalized or not. For this purpose, this study uses Severity Index for Suspected Arbovirus (SISA) dataset and implements various machine learning classification techniques with the aim of binary classification to detect the hospitalization status of a specific patient. Several neural networks, single classifiers, and ensemble supervised learning methods are selected as classifiers during the experiments. The best classification accuracy value is obtained by Random Forest (RF) model with 0.9908. This model has been shown to outperform many data mining techniques previously applied in prominent studies. This improved result leads to additional experiments with a different number of estimators when implementing RF. The outcome also improves the maximum classification performance up to 0.9926 using 25 estimators. The study reveals the effectiveness of ensemble models, especially bagging and boosting approaches, for Arbovirus suspected case analysis.}, publisher={Pamukkale Üniversitesi}