Knowledge Graph Recommendation Model Based on Feature Space Fusion

The existing recommendation model based on a knowledge graph simply integrates the behavior features in user–item bipartite and content graph. However, difference between two feature spaces is ignored. To solve this problem, paper presents new named space fusion (KGRFSF). Specifically, behavioral space, of users items are constructed by extracting from In related to extracted through attention mechanism graph, then vectors constructed. Finally, model, projected into same preference completed construct complete vector representations calculate similarity predict score user item. This applies presented public datasets fields music film. It can be found experimental results that KGRFSF effectively improve accuracy compared with models.