Knowledge graph attention network with path rotate encoding for recommendation

doi:10.1016/j.eswa.2025.130235

Knowledge graph attention network with path rotate encoding for recommendation

2025-11-04

0

PRE

AI

DOI：10.1016/j.eswa.2025.130235

Original

Original request for help

Save

Abstract

En 中文

• Discussed the limitations of knowledge graph-based recommendation systems. • A new model (KPRE) is proposed to address the limitations of existing methods. • KPRE can successfully combine path-based and propagation-based approaches. • KPRE can achieve advanced levels in CTR prediction and Top-K recommendation. • A series of experiments proved the effective and reasonable design of KPRE.

Knowledge graph attention network with path rotate encoding for recommendation

Abstract

Journal

Organization