<![CDATA[The high incidence of infectious and non-communicable diseases in Lhokseumawe City requires a data-driven approach for mapping distribution areas. This study designs a disease distribution clustering system using the Fuzzy Gustafson-Kessel (FGK) method and compares three fuzzy membership functions: linear, sigmoid, and Gaussian. FGK excels in handling non-uniform data distributions utilizing an adaptive covariance matrix based on Mahalanobis distance. The data comprised 38,063 cases from the Lhokseumawe City Health Office (2023–2024), aggregated into 4 sub-districts, and processed using z-score standardization. The clustering was implemented in a Python and MySQL web-based system and evaluated using the Xie-Beni validity index. The FGK method successfully grouped the sub-districts into 3 clusters: low, moderate, and high. For infectious diseases (converging at iteration 24), Blang Mangat formed the low cluster, Muara Dua and Muara Satu the moderate, and Banda Sakti the high. For non-communicable diseases (converging at iteration 9), Blang Mangat was low, Muara Satu moderate, while Banda Sakti and Muara Dua were high. The linear membership function proved optimal, yielding the lowest average Xie-Beni Index of 0.1316. In conclusion, the FGK method with a linear membership function effectively maps disease distribution, assisting the local Health Office in targeted policy planning and resource allocation.]]>
Copyrights © 2026
