Faqih Auliyaur Rohman, Harits Ar Rosyid, Agusta Rakhmat Taufani
Rapid advancements in technology, computing, and communication have significantly boosted internet capabilities, enabling diverse online activities. Gaming, initially perceived purely for entertainment, has evolved into professional careers and substantial business avenues, notably within e-commerce platforms focused on game sales and vouchers. However, e-commerce enterprises must continuously meet diverse user preferences and needs. This study proposes a game recommendation system leveraging content-based filtering coupled with the K-Means clustering algorithm. Utilizing Scikit-Learn, the optimal configuration for the K-Means algorithm identified is K-Means++ for centroid initialization, Elkan algorithm for cluster assignment, and a maximum iteration of 200, resulting in a high silhouette score of 0.712. This optimal setting effectively enhances recommendation accuracy within a web-based system. © 2025 IEEE.
Faculty of Engineering Universitas Negeri Malang, Informatics Engineering Study Program, Department of Electrical Engineering, Malang, Indonesia
