科技项目资助对论文产出绩效的作用效率测度研究

以人工智能领域论文和学者为研究对象，探讨科技项目资助对论文产出绩效的作用效率。运用负二项回归分析不同国家、领域资助种类、强度对论文引用和载文期刊水平的影响，结果发现，各项资助指标均可提升论文质量，但在不同地域和研究方向上的作用效果有所差异。运用倾向得分分层和线性回归，从发文、被引量和h指数角度测度资助对学者产出绩效提升的有效性，结果发现，资助能够显著提升学者产出力和影响力。通过Tobit回归测度资助项目数量和金额对不同年龄群体学者发文、被引量与变化率等指标的影响，结果发现，资助项目数量增加可整体提升学者产出绩效，资助金额增加更能扩大产出力和影响力的动态提升幅度。据此，提出资助改进策略和评审机制优化等相关建议。

Research on the Efficiency Measurement of the Effect of Science and Technology Project Funding on Output Performance of Paper

Based on the papers and scholars in the field of artificial intelligence explores the effect of the funding of science and technology projects on the output performance of paper. Specifically, the use of negative binomial regression analysis of different countries and sub-sectors of the type of funding, intensity, types of funding on the impact of citations and published journal level influence, found that the funding indicators can be positive on the quality of the paper,and there are differences in the effect of scholarship promotion in different regions and research directions.Using the tendency score method and linear regression models to measure the effectiveness of subsidized scholars' output performance improvement from the perspectives of the number of paper, cited and h index, find that funding can significantly improve the overall productivity and academic influence of scholars; and the number and the amount of the subsidized projects by Tobit regression measure have different effects on different different age group scholars,the results show that the increase of the number can improve the output performance of the scholars as a whole, and the increase of the amount can further expand the dynamic increase of output and influence. In view of this, relevant suggestions on funding improvement strategies and evaluation mechanism optimization are proposed.