A knowledge graph dataset for the study of Chinese scientific and technological literature

The knowledge graph dataset for the research of Chinese scientific and technological literature contains a total of 184078 nodes and 347360 edges, covering multiple disciplines such as computer science, medicine, agriculture, psychology, library and information science, and management. The dataset is designed based on a literature heterogeneous graph network, with node types including literature, author, institution, keyword, journal and research field. All entities are uniformly identified. The dataset supports multiple tasks such as literature recommendation, abstract generation, automatic journal selection, domain classification, graph representation learning, and influence dissemination. It is compatible with pre-trained language models like BERT and GPT, as well as graph neural network architectures such as GCN, GAT, and metapath2vec, and has excellent research and application value.