CBR: Category-Aware Cross-View Contrastive Learning Framework for Bundle Recommendation

Fuyong Xu, Zhenfang Zhu, Peiyu Liu, Ru Wang

Abstract: Bundle recommendation aims to suggest a series of items that users are interested in, and MealRec is a novel bundle recommendation dataset recently introduced. While current methods utilize user, bundle, and item interaction data to create user and bundle representations, the connection between item view and bundle view has not received sufficient attention. This study addresses this gap by examining the relationship between item view and bundle view. To capture interactive cooperative relations, we present the Category-aware Cross-view Contrastive learning framework (\(\text {C}^3\)BR). We participated in the workshop of BundleRS 2023 and the team name is Reset. Our experiments on MealRec dataset prove the superiority of \(\text {C}^3\)BR over baseline method.

External IDs:dblp:conf/dasfaa/XuZLW23a