An explainable recommendation framework based on an improved knowledge graph attention network with massive volumes of side information

Ryotaro Shimizu, Megumi Matsutani, Masayuki Goto

Abstract: Highlights•Reduced the computational cost of knowledge graph enabled recommendation.•Improved explainability while maintaining high recommendation accuracy.•Introduced how to apply knowledge graph embedding to marketing strategy planning.•Presented strengths of explainable recommendation, not just explainability.•Improved value and applicability of knowledge graph attention network to services.