<![CDATA[Classification is a systematic method of grouping data based on predefined analytical rules and principles. One of the classification methods employed in this study is the Modified K-Nearest Neighbors (MKNN) algorithm, which is recognized for its potential to achieve higher accuracy. MKNN is an extension of the traditional K-Nearest Neighbors (KNN) method, incorporating an additional ranking stage and a weighted voting mechanism using an alpha value of 0.5. The object of this study is stroke disease. In the medical context, stroke occurs due to a disruption of blood flow to the brain. Ischemic stroke is caused by the obstruction of blood vessels and is generally considered less severe, whereas hemorrhagic stroke results from the rupture of blood vessels and is categorized as a severe condition. Hospitals in Indonesia are required to provide prompt and accurate healthcare services, in accordance with Law Number 36 of 2009 concerning Health. Approximately 70% of stroke patients have a history of hypertension and heart disease, while around 87% experience psychological disorders such as anxiety and depression. Based on data obtained from Cut Meutia Regional General Hospital (RSUD Cut Meutia) in Lhokseumawe, the classification of stroke types is still performed manually through clinical observation. Therefore, this study proposes a stroke classification system based on the MKNN algorithm. The system utilizes 11 features and two diagnostic classes, namely ischemic stroke and hemorrhagic stroke, with a total of 100 medical record datasets divided into 80 training data and 20 testing data. Using a value of K = 5, the system achieved an average confidence accuracy of 81.19%, with a precision of 85.71%, recall of 80%, F1-score of 82.75%, and overall accuracy of 75%. The system was developed using the PHP programming language and a MySQL database.]]>
