Analisis Klaster Data Pasien Diabetes untuk Identifikasi Pola dan Karakteristik Pasien

Abstract

Diabetes is a significant health problem in Indonesia and the world. To understand the patterns and characteristics of diabetic patients, research was conducted by clustering the data of diabetic patients using the K-Means algorithm. The results of the analysis showed that there were two clusters, with cluster 1 consisting of 755 female patients aged 20-80 years and cluster 2 consisting of 404 male patients aged 40-90 years. The diagnosis of Non-insulin-dependent diabetes mellitus was the most common diagnosis in both clusters, followed by Rheumatoid arthritis in cluster 1 and Respiratory tuberculosis in cluster 2. BMI results in the "ideal" category had the highest frequency in both clusters, but the "less" category was more found in cluster 2. The unique variables in cluster 1 are M13.9 and I15, while the unique variables in cluster 2 are A15 and E11.6. In addition, the analysis of the two clusters shows that the Modopuro and Kebondalem sub-districts appear as the most sub-districts in the two clusters.

Perbandingan Algoritma Naïve Bayes dan K-Nearest Neighbors untuk Klasifikasi Metabolik Sindrom

“…Data Transformation adalah mengubah data untuk mendapatkan data yang lebih baik [21] dilakukan untuk mengubah data yang akan dianalisis sehingga dapat menghasilkan nilai yang lebih relevan [22].…”