CAJ | 학술논문

对2000年至2010年间某地不同类别高校科研综合投入与产出的面板数据采用Pearson简单相关分析和Logistics回归拟合，分析得出投入与产出的正相关性强的关键因子。同时，采用极大似然估计对数据分析并建立相应的科研绩效的高校分类模型，该模型较好地验证了2000年至2010年间不同类别高校的科研绩效实际；研究得出了不同类别高校科研绩效差异预测的点估计和95%Wald可信区间估计范围情况预测概率与观测值之间的关系，定量地解释了不同类别高校科研绩效的差异与发展趋势。模型对于合理评价科研绩效以及对于科研活动进行预测和调控，具有重要的现实意义。
대2000년지2010년간모지불동유별고교과연종합투입여산출적면판수거채용Pearson간단상관분석화Logistics회귀의합，분석득출투입여산출적정상관성강적관건인자。동시，채용겁대사연고계대수거분석병건립상응적과연적효적고교분류모형，해모형교호지험증료2000년지2010년간불동유별고교적과연적효실제；연구득출료불동유별고교과연적효차이예측적점고계화95%Wald가신구간고계범위정황예측개솔여관측치지간적관계，정량지해석료불동유별고교과연적효적차이여발전추세。모형대우합리평개과연적효이급대우과연활동진행예측화조공，구유중요적현실의의。
A Pearson Simple Correlation Analysis and a Logistics Regression Fitting have been employed to analyze the panel data about the overall inputs and outputs of R&D of various types of universities in a certain area from 2000 to 2010. The strong positive correlation of the key factors between inputs and outputs was obtained. Meanwhile, the Maxi-mum Likelihood Estimation was adopted to analyze the data for the establishment of a classified model of university R&D efficiency of various types,which appropriately verifies the actual efficiency of R&D from 2000 to 2010. The point esti-mation for the prediction of difference in efficiency of R&D between universities was concluded in the paper and the re-lation between observation and prediction probability of 95% Wald confidence interval estimation was also determined. The difference and development tendency of the university R&D efficiency of various types was explained quantification-ally. This modeling has practical significance for the rational evaluation of efficiency of R& D as well as the prediction and regulation of scientific research activity.