<![CDATA[Stroke is one of the leading non-communicable diseases that causes high mortality and long-term disability, making early risk prediction an important public health issue. In Indonesia, the increasing prevalence of stroke highlights the need for data-driven approaches to support early detection and prevention efforts. This study aims to compare the performance of Decision Tree and Neural Network algorithms in predicting stroke risk using health-related data. The research method employs a publicly available stroke prediction dataset obtained from Kaggle consisting of 5,111 records. Data preprocessing was conducted to handle missing values and prepare the dataset for modeling, followed by data splitting into training and testing sets using an 80:20 ratio. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. The results show that the Decision Tree model achieved higher overall accuracy of 80.72%, but demonstrated low recall for the stroke class. In contrast, the Neural Network model produced a lower accuracy of 69.37% but achieved a high recall of 82%, indicating better sensitivity in detecting stroke cases. These findings reveal a trade-off between overall accuracy and sensitivity in both models. It can be concluded that Neural Network is more suitable for stroke risk prediction when early detection is prioritized, while Decision Tree is preferable for achieving higher general classification accuracy.]]>
