<![CDATA[Hypertension is a serious concern because of its significant impact on public health, especially in the context of lifestyle changes and specific health conditions. One method for grouping patients based on complex clinical data is the Clustering method. This research type is quantitative, namely taking or collecting the necessary data and then analyzing it using the K-Medoids and K-Prototypes methods. The K-Medoids method is more resistant to outliers and noise than the K-Means method, which is more suitable for this research. The K-Prototypes method can handle mixed numerical and categorical data, effectively grouping hypertensive patients based on different variable categories. This research used the K-Medoids and K-Prototypes grouping methods to categorize patients into risk categories based on gender, age, family history of hypertension, smoking status, pulse rate, and increased systolic and diastolic blood pressure. The Elbow and Silhouette Coefficient methods were applied to evaluate the data and determine the optimal number of clusters for dividing patients into low-risk and high-risk hypertension groups. The analysis revealed that two clusters are the optimal solution. The clustering results show K-Medoids' superiority in grouping data with higher Silhouette Coefficient values compared to K-Prototypes. Overall, the K-Medoids and K-Prototypes algorithms can detect early hypertension risk by dividing patients into different risk groups. Although the clustering results are still weak, these two methods show potential in helping health institutions identify and treat hypertension risk in Indonesia.]]>
Copyrights © 2025
