OPTIMASI K-MEANS CLUSTERING PSO UNTUK PENENTUAN JUMLAH CLUSTER OPTIMAL PADA DATA KANKER PAYUDARA

Breast cancer is one of the most dangerous diseases and a leading cause of death among women worldwide. Clustering methods can assist in diagnosing breast cancer to determine the best course of treatment. K-Means is a widely used clustering algorithm known for its ability to handle large datasets efficiently with fast computational time. However, K-Means has a significant weakness: the number of clusters is determined randomly, resulting in suboptimal clustering outcomes. To overcome this limitation, Particle Swarm Optimization (PSO) is applied for automatic determination of the optimal number of clusters. PSO was selected due to its advantages, including requiring few parameters, ease of implementation, fast convergence, and low computational cost. This study uses the breast cancer dataset from the UCI Machine Learning Repository, consisting of 699 records and 10 attributes. The proposed PSO–K-Means method was evaluated using the Silhouette Coefficient and Davies-Bouldin Index. The results show that the optimal number of clusters is k = 2, achieving a Silhouette Coefficient of 0.92 and a Davies-Bouldin Index of 1.374. These results demonstrate that the PSO–K-Means method significantly outperforms standard K-Means by directly producing optimal clustering results without the need for conducting repeated experiments.