<![CDATA[Hypertension and diabetes are non-communicable diseases whose prevalence continues to increase, making them major public health concerns in Indonesia. Perawas Community Health Center has a substantial amount of patient laboratory data; however, the data are still used mainly as examination archives. This study aims to develop a hypertension and diabetes prediction system using the Random Forest algorithm combined with the Synthetic Minority Oversampling Technique (SMOTE). The dataset consists of 1,722 patient records and 14 optimal features selected through feature selection. The test results show that the Decision Tree model achieved an accuracy of 81.74% before SMOTE and 73.62% after SMOTE, while Random Forest achieved an accuracy of 84.35% before SMOTE and 81.74% after SMOTE. The Random Forest model with SMOTE was then implemented in a web-based application to support the disease prediction process. The findings indicate that the Random Forest algorithm and SMOTE can support rapid, data-driven early detection of hypertension and diabetes based on patient laboratory data.]]>
Copyrights © 2026
