Social-Aware Movie Recommendation Using Graph Convolution Network

Effective recommendation systems that tailor content are now necessary due to the growth of online multimedia platforms. Many current approaches leverage social network information to improve recommendation accuracy, even though Graph Convolutional Networks (GCNs) have demonstrated strong performance in capturing high-order user-item associations. We present a Social-Aware Graph Convolution Network (SA-GCN) for movie selection in this research, which combines social relationships and user-item interactions into a single graph. To mitigate data sparsity and improve representation learning, SA-GCN simultaneously propagates preference signals across social and interaction graphs. We compare the SA-GCN against six baselines, including LightGCN and DiffNet++, using four datasets: MovieLens-1M, Douban Movie, Last.fm, and Epinions. Our findings show that the proposed SA-GCN performs better than all baselines throughout. The results validate the effectiveness of social-aware message passing in GCN-based recommendation and highlight its potential for broader applications in social network-driven personalization tasks.