Leveraging Member-Group Relations via Multi-View Graph Filtering for Effective Group Recommendation

Kim, Chae-Hyun, Choi, Yoon-Ryung, Park, Jin-Duk, Shin, Won-Yong

arXiv.org Artificial Intelligence 

Group recommendation aims at providing optimized recommendations tailored to diverse groups, enabling groups to enjoy appropriate items. On the other hand, most existing group recommendation methods are built upon deep neural network (DNN) architectures designed to capture the intricate relationships between member-level and group-level interactions. While these DNN-based approaches have proven their effectiveness, they require complex and expensive training procedures to incorporate group-level interactions in addition to member-level interactions. To overcome such limitations, we introduce Group-GF, a new approach for extremely fast recommendations of items to each group via multi-view graph filtering (GF) that offers a holistic view of complex member-group dynamics, without the need for costly model training. Specifically, in Group-GF, we first construct three item similarity graphs manifesting different viewpoints for GF. Then, we discover a distinct polynomial graph filter for each similarity graph and judiciously aggregate the three graph filters. Extensive experiments demonstrate the effectiveness of Group-GF in terms of significantly reducing runtime and achieving state-of-the-art recommendation accuracy.

Duplicate Docs Excel Report

Title
None found

Similar Docs  Excel Report  more

TitleSimilaritySource
None found