KLASTERISASI DATA REKAM MEDIS PASIEN RAWAT INAP MENGGUNAKAN METODE K-MEANS PADA RUMAH SAKIT DR. TADJUDDIN CHALID

Abstract

Electronic Medical Records (EMR) in hospitals generate large volumes of data that have not been optimally utilized for management decision-making. This study applies the K-Means algorithm to cluster inpatient medical record data at Dr. Tadjuddin Chalid Hospital and visualizes the results through an interactive dashboard. The dataset covers 1,848 patient records with attributes of diagnosis, residential area, age, and gender. The research stages include preprocessing, data transformation, K-Means clustering, and validation using the Silhouette Index. The results show that the optimal model is formed at K=2 with a Silhouette Index value of 0.8030 (strong structure). Cluster 1 is dominated by local patients (close proximity) aged adults to elderly with primary diagnoses of typhoid fever and dyspepsia. Conversely, cluster 2 is dominated by pediatric patients from outside Makassar City with a primary diagnosis of neonatal jaundice. This study also successfully developed a web-based dashboard to visualize the clustering results interactively to support management in data-driven decision-making.