Multi-Modal Knowledge Graph System with Embedding Module

Recommendation system is to predict appropriate Point of Interest (POIs) which have not visited before the users provides its historical records of check-ins. The implementation of knowledge graph system resolves the heterogeneity in auxiliary data. Though, previous knowledge graph system about the POI recommendation makes the utilization of relationship among POIs and users for context data to recommendation system and neglected the deep relation taken attribute integration of POIs itself. Many researches developed the Knowledge Graphs (KGs) based system through leveraging the significant auxiliary data. Many researches ignored the difference of data types in multi-modal. In this article, developed the Multi-modal Knowledge Graph Embedding Module (MKGEM) for enhancing the recommender system through knowledge of multi-modal. In MKGEM, the multi-modal graph attention method is evaluated to data propagation and next used the resulted aggregation representation of embedding for knowledge graph system. The developed MKGEM showed effective performance while compared to existing techniques like Collaborative Knowledge base Embedding (CKE).