Cross-view contrastive representation learning on meta-path induced graphs with node features for bundle recommendation

Puntos clave

• Improved accuracy in bundle recommendation is achieved through novel contrastive representation learning.
• Key metric shows that using meta-path induced graphs enhances recommendation quality by engaging node features.
• Cross-view contrastive representation technique effectively utilizes meta-paths to analyze graph structures.
• Results support the notion that leveraging meta-paths offers promising avenues for future recommendation systems.