Visualization analysis of research hotspots and frontier development of knowledge graph construction
Zhaodi Wang, Lei Wang, Gang Wang · Research Square · 2022
Abstract Knowledge graph construction is one of the hot issues in artificial intelligence, and it is the core and key to knowledge graph research. Through Citespace (a statistical software) and bibliometric analysis, we draw a visual knowledge graph based on relevant literature published from 2012 to 2021, and analyze it from multiple dimensions including annual published articles, institutions, research hotspots, and research frontiers. The following conclusions are obtained. First, the number of research results is gradually increasing, but the research institutions are scattered. Second, knowledge graph construction has been studied from diversified perspectives, but the concept connotation and technical boundary are vague. Third, the construction technology of each link is constantly developing, but its accuracy, basic methods and theories need to be further improved. Fourth, large-scale universal multilingual knowledge extraction and application, together with natural language processing, has become the most concerned issue in the current knowledge graph construction. However, in large-scale application in special fields, the research on knowledge graph evaluation technology is inadequate, and the development of educational technology based on knowledge graph is slightly insufficient. In the future, priority should be given to developing the application of knowledge graph in large-scale domestic fields, strengthening international cooperation and exchange in technology research & development, and developing knowledge graph system engineering conception. It is also important to strengthen the research on multi-domain application technology and the humanistic research of knowledge graph construction.