中国科技期刊研究 ›› 2017, Vol. 28 ›› Issue (4): 344-349.doi: 10.11946/cjstp.201701190046

• 评价与分析 • 上一篇    下一篇

基于论文作者特征的期刊影响力预测

李秀霞1),邵作运2)   

  1. 1) 曲阜师范大学传媒学院,山东省日照市烟台路80号 276826
    2) 曲阜师范大学日照校区图书馆,山东省日照市烟台路80号 276826
  • 收稿日期:2017-01-09 修回日期:2017-03-13 出版日期:2017-04-15 发布日期:2017-04-15
  • 作者简介:李秀霞(ORCID:0000-0002-3492-4768),副教授,硕士生导师,E-mail: zyshao@126.com|邵作运,硕士,馆员
  • 基金资助:
    国家社会科学基金(16BTQ074)

Prediction of journal influence based on the features of authors

LI Xiuxia1),SHAO Zuoyun2)   

  1. 1) School of Communication, Qufu Normal University, 80 Yantai Road, Rizhao 276826, China
    2) Library of Qufu Normal University, 80 Yantai Road, Rizhao 276826, China
  • Received:2017-01-09 Revised:2017-03-13 Online:2017-04-15 Published:2017-04-15

摘要:

【目的】 构建一组反映期刊内部特征信息的作者特征空间向量,以拓展期刊影响力分析方法。【方法】 以图书情报学领域18种核心期刊2011年第1期的380篇论文为研究对象,选取其中3/4的文献为训练样本,构建基于作者特征的期刊影响力预测模型,以剩下的1/4论文为测试样本,检验预测模型的有效性。【结果】 实验发现,期刊影响力预测模型与4年后对应期刊的影响因子具有较好的吻合度。【结论】 说明由期刊作者特征研究期刊影响力是可行的,为研究期刊影响力提供了一种新方法。

关键词: 作者特征, 期刊影响力, 相关分析, 曲线回归

Abstract:

[Purposes] In order to extend the evaluation method of journal influence, this paper aims to construct a set of feature space vectors, which can reflect the internal characteristics of journals.[Methods] We took 380 documents from the first phase of 18 core journals in the field of Library and Information Science in 2011 as the research object, obtained the prediction model of journal influence based on characteristics of authors by selecting 3/4 of the 380 literature as training samples based on correlation analysis and regression analysis, and checked the validity for the prediction model with the remaining 1/4 of the literatures.[Findings] The experimental result indicates that the prediction model of journal influence has a good agreement with the impact factors of journals after four years.[Conclusions] This research shows that it is feasible to study the influence of journals by the characteristics of journal authors. This study provides a new method to explore the influence of journals.

Key words: Author features, Journal influence, Correlation analysis, Curvilinear regression