基金资助的科学产出评估和预测方法研究OA北大核心CHSSCDCSSCICSTPCD

[Purpose/Significance]The evaluation and prediction of scientific output of fund projects can help opti-mize the effectiveness of fund support and improve the quality of scientific research.It has certain theoretical and practical guidance significance for the optimization of layout direction and resource allocation of scientific research management de-partments.[Method/Process]Firstly,based on the LDA model,identified the research topics contained in the fund and its supporting papers;Then,five indicators were proposed:the amount of topic funding,the duration of topic funding,the popularity of topics,the quality of topics,and the influence of topics.A comprehensive evaluation model for science output topics supported by the fund was designed;On this basis,a scientific output topic prediction model was proposed based on deep learning algorithms;Finally,the paper took the data of the papers funded by the Department of Social and Economic Sciences of the National Science Foundation and its output as an example.[Result/Conclusion]Through empiri-cal research,it was found that the high-quality scientific output topics of the Department of Social and Economic Affairs of the National Science Foundation of the United States mainly include policy and environment,scientific research develop-ment and practice,and health and social development,verifying the feasibility and effectiveness of the method.[Innova-tion/Limitations]This study has certain theoretical and practical guidance significance for optimizing the effectiveness of fund funding,improving the quality of scientific research,optimizing the layout direction of scientific research management departments,and optimizing resource allocation.