<![CDATA[Monkeypox is a contagious disease that requires prompt and accurate handling, particularly in the diagnostic process. However, identifying symptomps manually often takes time and is prone to error. In response to this challenge, this study aims to develop a machine learning based classification model to support a more efficient diagnosis process. This research applies two machine learning algorithms XGBoost regression and Random Forest regression to classify patients as infected or uninfected with monkeypox based on clinical symptoms. The study focuses on assessing how well each algorithm can distinguish between positive and negative cases, especially when dealing with imbalanced data or overlapping features. The dataset used consists of 25.000 entries sourced from Kaggle, each containing clinical indicators related to monkeypox. Before modeling, the data underwent exploratory data analysis (EDA) and preprocessing, including handling missing values. Furthermore, cross-validation and parameter tuning techniques were implemented to optimize model performance. The results indicate that XGBoost outperformed Random Forest, achieving 68% accuracy, 69% precision, 89% recall, and a 78% F1-score. In contrast, Random Forest yielded slightly lower scores. Both models were evaluated using the ROC curve, where each reached an AUC values of 0.60. This suggests that wile both models show potential, their ability to clearly distinguish between classes positive and negative remains limited and can be improves in future work.]]>
Copyrights © 2025
