科研评价的统计模型

Statistical Model of Scientific Research Evaluation

传统的科研评价指标(如h指数、g指数)没有考虑到学科差异性,不同学科的科研人员之间没有可比性。同时,通过统计分析后,我们发现论文的被引量与发表的时间成正相关,不能简单地以被引频次衡量不同时期的论文。因此,为了消除根据学科差异性和论文发表时间带来的影响,我们建立了新的模型,对h指数和g指数进行了修正,同时提出了新的相对水平指标:REL指数作为对修正后的h指数和g指数的补充和拓展。

There are numerous traditional indexes （such as h- index and g-index） to quantify an individual＇s scientific research output. However, most of them cannot characterize the scientific output of researchers in different disciplines reasonably due to the fact that they ignore the gaps between different disciplines. Besides, after statistical analysis, we found that citations are related to the published time and it is unreasonable to measure the articles which are published in different time by citations. As a result, in order to eliminate the impact of disciplines and published time, we build a statistical model to eoneet the h-index as well as g-index and put forward a new index： REL-index, as the supplement and development of h-index and g-index.