Research On The Research Frontier Topic Identification And Evolution Law Based On NSF Data

In the era of big data,it is of great strategic significance to realize the development strategy of science and technology innovation and improve the influence of international scientific research in the process of grasping the research front and the development trend of science and technology in the world as the key point for government science and technology management and scientific research institutions to seize the opportunities.How to identify the research front from the mass of information science and technology,and promote scientific and technological innovation has become one of the key and hotspots of information science research.In the field of research front identification,it has gradually formed more mature theories and methods,such as citation analysis and keyword analysis,which are widely used in the prediction of different countries and disciplines.However,there are some problems between the method of citation analysis and the analysis method of the subject words,such as the citation delay problem and analysis of content relative to the lack of correlation analysis,which limiting the timeliness and accuracy of research front recognition prediction.With the development of information technology,the researchers begin to use the visualization technology and in-depth analysis of the evolution of the text topic.It focuses on the split,integration and other complex evolution process from the time dimension of the research front topic in order to be able to improve the method of citation and text content analysis.In this paper,the fund project text of the National Science Foundation(National Science,Foundation,NSF)as the data source,using topic identification,complex network analysis and visualization analysis technique to identify research front topic and its development trend.There are three steps in the research of the research front topic identification: topic identification,the research front topic determination and topic evolution analysis,specifically,the main research is as follows:(1)topic identification based on the PLDA model.Detecting research topic in the text based on the PLDA model,and then build the topic correlation network by visualization technology to identify the association of topics;(2)research on front topic determination based on the multi-dimension topic evolution analysis model.From three dimensions the topic strength,topic structure and topic content to build the topic evolution analysis model,to analysis the complex process of topic evolution,determine the research front topic;(3)research front topic evolution analysis based on topic evolution visualization.This paper designs three kinds of innovative scientific knowledge maps,from three dimensions of the topic strength,topic structure and topic content to do visual analysis,help to find the research front topic evolution path and predict its development trend.The results show that research front topic identification and visual analysis can effectively identify the research front topic and its evolution in the fund project data.It can help to analyze the complicated process of researching the evolution of front topics and provide information based on scientific research to improve efficiency and promote relevant areas of scientific and technological innovation.But this paper only uses the fund project data that has some limitations,and needs to be expanded from the data sources.