图书情报工作 ›› 2017, Vol. 61 ›› Issue (17): 122-130.DOI: 10.13266/j.issn.0252-3116.2017.17.016

• 情报研究 • 上一篇    下一篇

引文分布视角下期刊评价的P指数研究

陈卫静, 张宇娥, 蔺梅芳   

  1. 电子科技大学图书馆 成都 611731
  • 收稿日期:2017-05-08 修回日期:2017-07-13 出版日期:2017-09-05 发布日期:2017-09-05
  • 作者简介:陈卫静(ORCID:0000-0003-4380-9638),馆员,硕士,E-mail:chenwj@uestc.edu.cn;张宇娥,副研究馆员,硕士;蔺梅芳,馆员,硕士。
  • 基金资助:
    本文系电子科技大学中央高校基本科研业务费项目(基础研究类) "基于ESI的中国高校学科发展评价模型研究" (项目编号:ZYGX2015J178)和电子科技大学图书馆项目"学术成果影响力评价方法及体系构建研究"(项目编号:LS2015Z001)研究成果之一。

Research on P-index:Journal Evaluation Under the Context of Citation Distribution

Chen Weijing, Zhang Yu'e, Lin Meifang   

  1. Library of University of Electronic Science and Technology of China, Chengdu 611731
  • Received:2017-05-08 Revised:2017-07-13 Online:2017-09-05 Published:2017-09-05

摘要: [目的/意义]针对P指数不能反映评价对象引文分布的P缺陷,在P指数研究的基础上,提出基于引文分布的指数计算方法。[方法/过程]引入发文-引文分布曲线和学术矩阵表征期刊的引文分布,利用期刊评价的非参数统计方法,通过将引文数据进行分区划分,基于P核和P尾内有引用的论文,采用期刊的平均引文率代表零引用论文的影响,重新定义P指数的计算方法,提出Pnew指数。为避免学科差异对期刊评价的影响,以仅属于Web of Science数据库中电子电气工程类目下有5年影响因子的53种期刊为研究对象,分别统计期刊的影响因子、5年影响因子、H指数、P指数、Pnew指数等指标,并进行相关性及差异性分析,以验证Pnew指数评价的科学性和有效性。[结果/结论]结果表明,Pnew指数与H指数、P指数等多个关键指标的相关性较好。与P指数相比,Pnew指数既延续了其高灵敏度与高区分度的特点,又对P指数相差无几但引用分布大不相同的期刊有更好的显示度和辨识度,而且能够矫正引文分布对P指数评价的偏差。

关键词: 期刊评价, 引文分布, 学术矩阵, P指数, Pnew指数

Abstract: [Purpose/significance] Taking account of the difference of citation distribution, this paper presents a new calculation method for P-index, that is Pnew-index, to overcome the disadvantage of P-index which cannot reflect the distribution of citation.[Method/process] This paper introduced the rank-citation curve and the academic matrix to characterize the citation distribution of journals. By using the nonparametric statistical method, this paper divided the citation data into different zones, and reconstructed the calculation method of P-index based on the P-core and P-tail papers which had citations. In order to avoid the effect of subject difference, we selected 53 journals which had 5-year impact factors from the category of electrical and electronic engineering in the Web of Science database. Then, different indicators were calculated for each journal, including impact factors, 5-year impact factors, H-index, P-index, Pnew-index, Pc-index and Pt-index. This paper also analyzed the correlation of these indicators to verify the scientificity and effectiveness of the Pnew-index. [Result/conclusion] The result shows that the Pnew-index has strong correlation with H-index, P-index and other key indexes. Comparing with the P-index, the Pnew-index extends its advantage of high sensitivity and strong discrimination; has better visibility and recognition for the journals which have almost the same P-index but great different citation distribution; and can correct the deviation of P-index caused by citation distribution.

Key words: journal evaluation, citation distribution, academic matrix P-index, Pnew-index

中图分类号: