当前位置: X-MOL 学术Scientometrics › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Ranking of author assessment parameters using Logistic Regression
Scientometrics ( IF 3.5 ) Pub Date : 2020-11-20 , DOI: 10.1007/s11192-020-03769-y
Muhammad Usman , Ghulam Mustafa , Muhammad Tanvir Afzal

The renowned international scientific societies nominate researchers for awards based on qualitative judgments every year. Qualitative judgment uses subjective assessments based on information that is not quantifiable. The way of assessing the quality of the work has not been established or disclosed, nor do we have any qualitative evaluation criteria. We can assess the quality of the researcher's work by mapping the quantitative parameters to qualitative judgments. To date, the scientific community has presented more than 50 research assessment quantitative parameters, including publication count, citation count, h-index, and its variants. The contemporary state-of-the-art in authors ranking does not determine the best parameter that effectively maps on experts' qualitative evaluation. Moreover, these parameters have been evaluated by using same scenarios. In such scenarios, the value and effect of each parameter over the others are complicated to ascertain. Therefore, they must be assessed in inequitable scenarios. The purpose of this research is to identify the significant parameters that map on qualitative judgments of international scientific societies in Civil Engineering (CE) for award nominations. We will identify the rank of author assessment parameters, which includes published papers, citations, No of years since 1st publication, citations in h-core, authors/paper, citations/paper, citations/year, h-index, g-index, hg-index, A-index, R-index, e-index, and f-index. We have evaluated these parameters on the dataset from the discipline of Civil Engineering (CE). The data set contains 250 non-award winners and 250 award winners from prestigious scientific societies of CE. The h-index and its variants have been ranked based on their effectiveness for awardees using Logistic Regression. The award-winning researchers have less number of average authors/paper than the non-awardees. The authors/paper has achieved the highest effectiveness of 67% for awardees. Furthermore, we have also analyzed the ratio of awardees in the ranked list of 50, 100, and 150 researchers by author assessment parameters. The authors/papers have outperformed all other indices by elevating 62% and 66% of the award recipients in its ranked list of 100 and 150 researchers. In the ranked list of 50 researchers, publications elevate 54% awardees, and Authors/papers achieved the second-highest elevation score of awardees of 50%.

中文翻译:

使用逻辑回归对作者评估参数进行排名

著名的国际科学协会每年都会根据定性判断提名研究人员获奖。定性判断使用基于不可量化信息的主观评估。评估作品质量的方式尚未建立或公开,我们也没有任何定性的评估标准。我们可以通过将定量参数映射到定性判断来评估研究人员工作的质量。迄今为止,科学界已经提出了 50 多个研究评估量化参数,包括出版物计数、引用计数、h-index 及其变体。当代最先进的作者排名并不能确定有效映射专家定性评估的最佳参数。而且,这些参数已通过使用相同的场景进行评估。在这种情况下,要确定每个参数相对于其他参数的值和影响是很复杂的。因此,必须在不公平的情况下对它们进行评估。本研究的目的是确定影响国际土木工程 (CE) 科学学会定性判断的重要参数,以供提名。我们将确定作者评估参数的排名,其中包括已发表论文、引文、自第一次发表以来的年数、h-core 中的引文、作者/论文、引文/论文、引文/年、h-index、g-index、 hg-index、A-index、R-index、e-index 和 f-index。我们已经在土木工程 (CE) 学科的数据集上评估了这些参数。该数据集包含来自 CE 著名科学学会的 250 名非获奖者和 250 名获奖者。h 指数及其变体已使用逻辑回归根据其对获奖者的有效性进行排名。获奖研究人员的平均作者/论文数量少于未获奖者。作者/论文对获奖者的有效性达到了 67%。此外,我们还通过作者评估参数分析了 50、100 和 150 名研究人员的排名列表中的获奖者比例。作者/论文在其 100 名和 150 名研究人员的排名中分别提升了 62% 和 66% 的获奖者,其表现优于所有其他指数。在排名 50 的研究人员名单中,出版物提升了 54% 的获奖者,作者/论文获得了 50% 的获奖者排名第二高的分数。
更新日期:2020-11-20
down
wechat
bug