为了分析学术期刊评价指标与评价结果的数据分布关系，基于JCR2012物理学期刊数据，采用偏度、峰度、中位数均值比、离散系数、集中度、幂指数、拟合优度等指标，综合比较了评价指标数据分布与熵权法、因子分析、TOPSIS三种评价方法评价结果数据分布的关系。研究发现，特征因子、总被引频次的数据不均衡水平最大，而被引半衰期的数据不均衡水平相对较好；多属性评价会显著降低期刊评价指标的数据不均衡水平；TOPSIS法降低数据不均衡的效果最好，熵权法次之，因子分析法最差；评价方法的数据均衡水平也可以成为评价方法的选择标准。%This paper based bases on the index data of JCR2012 Physics journal bibliometrics, . we using used some indexes such as skewness, kurtosis, median average ratio, dispersion coefficient, concentration, exponential, etc. We compared the evaluation results re-lationship of data distribution between evaluation index data distribution and other three evaluation methods including entropy weight meth-od, factor analysis and TOPSIS comprehensively. Our study found that:for the data imbalance level, characteristic factor and total cited frequency were the maximum, While the cited half-life was relatively good. Multi attribute evaluation will significantly reduce the data imbalance level of journal evaluation index. The method of TOPSIS decreased the data imbalanced had the best effect. Entropy method fol-lowed, while the factor analysis was close to the imbalance level of original evaluation index. So the data balance level may also be the se-lection criteria of evaluation methods.
展开▼
