RecKG-Web: A web-based interactive visualization tool for standardized knowledge graphs in recommender systems

초록

Knowledge graphs (KG) enhance recommender systems (RS) by addressing challenges like the long-tail and cold-start problems through rich semantic relationships. However, inconsistent metadata across RS datasets limits interoperability. To solve this, we present RecKG-Web, a user-friendly application that standardizes and integrates heterogeneous datasets using the RecKG format. It supports data upload, mapping, graph-based visualization, and dataset export, enabling cross-domain recommendations and improving accessibility for researchers and practitioners. RecKG-Web bridges the gap between KG theory and practice, advancing KG-aware RS development.

키워드