<![CDATA[This study aims to segment smartphone products based on the variables of Price, Rating, and Number of Reviews using the K-Means and K-Medoids methods. The dataset used consists of 400 smartphone products that have undergone data normalization as part of the preprocessing stage. The number of clusters was determined using an internal evaluation method, and the optimal number of clusters was found to be four (K=4). The clustering results show that both methods are capable of forming significantly different product group characteristics based on a combination of price level, user rating quality, and review intensity. The K-Means method produces a more structured cluster separation based on centroid values and is effective in representing the average data distribution. Meanwhile, K-Medoids demonstrate better resilience against outliers because cluster centers are represented by actual objects (medoids), making them more stable on heterogeneous data. Based on a comparative analysis of the methods’ characteristics and cluster evaluation results, K-Medoids demonstrates more robust performance for datasets with significant price variation. The findings of this study can serve as a basis for decision-making in marketing strategies and product clustering on e-commerce platforms.]]>
Copyrights © 2026
