Knowledge Graph and GNN-Based News Recommendation Algorithm With Edge Computing Support

Abstract

The current news information from different media websites has posed a serious problem, i.e., it is very difficult to obtain the satisfactory news contents from the measured data information. There have been some researches on news recommendation to improve the experience of users. In spite of this, they always need the further improvement because the news information has showed the explosive increasing way. Therefore, this paper studies knowledge graph and graph neural network (GNN) based news recommendation algorithm with edge computing consideration. At first, the knowledge graph is used for the knowledge extraction. Then, GNN is used to train the extracted features to complete the news recommendation algorithm. Finally, the edge computing is used to offload the high volumes of traffic to the edge server for the news recommendation computation. Compared with two baselines, the proposed algorithm is more efficient, increasing accuracy rate by 2.73% and 9.94% respectively, and decreasing response time by 84.27% and 87.58 respectively.