Sistem Rekomendasi Film Menggunakan Data User-End dan Knowledge Graph Convolutional Network pada Dataset MovieLens 1 M

Yanuar, Muhammad Rizki
Umbara, Fajri Rakhmat
-, Agus Komarudin

JURIKOM (Jurnal Riset Komputer)
Vol. 12 No. 4 (2025): Agustus 2025

Traditional recommendation systems such as Collaborative Filtering and Content-Based Filtering often fail to provide relevant recommendations due to their limitations in handling sparsity and cold-start problems. This study proposes a Knowledge Graph Convolutional Network (KGCN) model enriched with user demographic data from the MovieLens 1M dataset to address these issues. The primary focus of the research is to demonstrate that the Importance Sampling technique is significantly superior to Uniform Sampling in effectively training the model. After hyperparameter tuning, the optimal model configuration achieved peak performance with an AUC score of 0.8798 and NDCG@10 of 0.9719. These results demonstrate that the proposed approach is effective in building an accurate, personalised recommendation system capable of addressing sparsity and cold-start issues.