基于关联数据的科研机构评价指标权重计算方法研究

Research on Weigh Computing Method of Evaluation Indicators of Scientific Research Institution Based on LOD

[目的 /意义]科研评价中指标权重计算的合理性将直接影响科研机构评价结果的客观性和准确性,本文提出利用信息论方法来计算指标权重,为指标权重计算提供一种新的思路。[方法 /过程]基于DBpedia数据集,利用信息论方法计算出的指标权重和上海交通大学世界大学排行榜已有的指标权重,同时对榜单前100名大学机构进行排名,并将排名结果进行对比分析。[结果 /结论]实验发现利用本文提出的权重计算方法得出的机构得分结果与上海交通大学已有指标权重的得分结果皮尔逊相关性为0.980,斯皮尔曼相关性为0.939,并且其排名顺序和上海交通大学给出的排名顺序皮尔逊相关性和斯皮尔曼相关性均为0.939。以上两个排名结果的得分相关性和排名相关性极强,证明本研究中关联数据的权重计算方法的有效性。

[Purpose/significance] The rationality of calculating the indicators weight in the scientific research evaluation directly affects the accuracy and objectivity of evaluation results of scientific research institution. This paper proposes a new method for computing the indicators weight by the information theory.[Method/process] Based on the DBpedia set and information theory, it computes the indicator weight and uses the existing indicator weight of Shanghai Jiaotong University＇s World University Rankings, makes a ranking experiment on the list of top 100 universities and makes a comparative analysis on the two kinds of ranking results.[Result/conclusion] The Pearson correlation between the score results with the proposed weight computing method and of Shanghai Jiaotong University＇s existing weights is 0.980, Spearman correlation is 0.939. The Pearson correlation and Spearman correlation between its ranking order and Shanghai Jiaotong University＇s are both 0.939. The score correlation and rank correlation of above two ranking results are very strong, and prove the validity of the weight computing method based on the linked data in this study.