Research Hotspot Prediction and Regular Evolutionary Pattern Identification Based on NSFC Grants Using NMF and Semantic Retrieval

Abstract

Analyzing the research hotspots and regular evolutionary patterns of R&D projects can help researchers find potential information. This paper considers the titles of the National Natural Science Foundation of China (NSFC) grants that have been awarded in the past 20 years as the research objects. First, we analyze the number and funding amounts of project grants for each department over the past 20 years. Second, we propose a topic discovery method that is based on nonnegative matrix factorization and further propose a keyword scoring method that is based on semantic retrieval, from which we can discover the regular evolutionary patterns of research hotspots. Finally, we conduct experiments on datasets on grants in the Department of Information Science. To identify research characteristics, trends and prospects, we explore the regular evolutionary patterns of research hotspots via experiments using three methods from multiple perspectives: word cloud, topic corresponded to keywords display and topic evolution display. The results of the experimental studies demonstrate that researchers can keep abreast of the main content and hotspots in the research field via hot spot discovery and the identification of regular evolutionary pattern of NSFC grants. This study can also help the government improve the allocation of scientific and technological resources and help decision makers make scientific decisions.

Keywords

• NSFC
• nonnegative matrix factorization
• hotspot prediction
• regular evolutionary pattern
• semantic association