The graph-based collaborative filtering method has shown significant application value in recommendation systems, as it models user-item preferences by constructing a user-item interaction graph.