Research On Emerging Topics Recognition And Development Trend Prediction Of Government Funded Projects Based On Text Mining

With the development of computers and the coming of the era of big data,the emerging topics representing the direction of future disciplines have been drawn attention from all levels of government and technology policymakers.At the same time,according to the development direction of frontiers,the key to the scientific and technological institutions to seize the scientific research institutions is to make the direction of industrial development and deploy the structure scientifically.In the field of emerging topic research,mature theories and research methods have been gradually formed,such as citation analysis method based on paper analysis,common cited method and other traditional information science research methods,so as to predict and analyze the new trend of disciplinary development.At the same time,based on the research method of the paper,because the analysis data source is the data source with time lag,no matter how the research and detection methods are modified or improved,it can not change the essential attribute of the time lag of the paper.With the development of computer technology and Natural Language Processing technology,using Natural Language Processing technology,in-depth text content,research topic of deep mining of text content,identifying research focus and emphasis of research and deployment,achieved fruitful results,the traditional method is to a certain extent,improve the scientific and accurate identification of emerging themes,but content analysis of relative isolation and lack of deep semantic relations,single data source types,restrict the development of new technology and improve the related topic detection.In this paper,we use the National Science Foundation(National Science,Foundation,NSF)of the government funded projects for the text analysis of the data source,using the topic model,machine learning,visual analysis,identify emerging themes contained,so as to analyze the research topic in the future.This research can be divided into three steps: topic probability recognition,government funded project emerging topic recognition,and emerging topic prediction analysis.Specifically,the main points are:(1)topic probability recognition model based on PLDA model.Based on the probability model,the text content is identified,and the topic words and related weight distributions in the government funded project texts are identified.(2)new topic identification of government funded projects.This paper analyzes the subsidy intensity,the amount of subsidy and thematic strength of the government funded projects,and establishes a set of new topic detection formulas based on the text analysis of the government funded projects,and analyzes the development and evolution of the themes in the text.(3)new topic prediction analysis and visualization based on machine learning model.In this part,we will use machine learning theory and method to process sliding window for data,establish time series model,predict and analyze future subject development themes,and visualize and analyze emerging themes with visualization technology.The experimental results show that the proposed new topic identification,prediction and visualization analysis technique can effectively identify with emerging topics in government funded projects in and to analyze and forecast its future development,so as to provide decision-making support and reference for the science and technology policy,to provide theoretical and technical support for our country and discipline adjustment the focus of the research on the future direction of development and deployment,research and innovation,to provide scientific research efficiency.Meanwhile,we use the three thematic features of the government funded project to identify and construct the new topic detection model.In the future,we will further analyze and verify the related features,and complete the new topic detection system.