图书情报工作

图书情报工作 >

2013 , Vol. 57 >Issue 04: 120 - 124,137

DOI: https://doi.org/10.7536/j.issn.0252-3116.2013.04.021

情报研究

基于主题模型的文献引用贡献分析

  • 张金松 ,
  • 陈燕 ,
  • 刘晓钟
展开
  • 1. 大连海事大学交通运输管理学院;
    2. 美国印第安纳大学布鲁明顿分校
张金松,大连海事大学交通运输管理学院博士研究生,E-mail:zhangjinsong85@163.com;陈燕,大连海事大学交通运输管理学院教授;刘晓钟,美国印第安纳大学布鲁明顿分校副教授。

收稿日期: 2012-11-23

  修回日期: 2013-01-09

  网络出版日期: 2013-02-20

收起

Literature Citation Contribution Analysis Based on the Topic Model

  • Zhang Jinsong ,
  • Chen Yan ,
  • Liu Xiaozhong
Expand
  • 1. Dalian Maritime University, Dalian 116026;
    2. Indiana University Bloomington, Indiana USA 47408

Received date: 2012-11-23

  Revised date: 2013-01-09

  Online published: 2013-02-20

Fold

摘要

传统引文分析方法中,文献间的相互关系通常由引用关系决定,也就是说,如果文献A引用文献B,则证明B对A有一定的贡献,然而具体的贡献值与引用原因却很难进行界定。采用主题模型的方法,将原著、引文、被引文献看作是主题模型上的概率分布,通过全文抽取的方法,对引用的原因以及引文贡献值进行分析。首先介绍研究背景与研究意义,并对基本概念进行阐述;然后介绍引文抽取方法、 利用Labeled-LDA模型建立主题模型方法等;最后通过实验部分建立基于不同主题的文献引用网络图,并利用工具使其可视化表示。

关键词： Labeled-LDA; 主题模型; 引文分析; 全文抽取

本文引用格式

张金松 , 陈燕 , 刘晓钟 . 基于主题模型的文献引用贡献分析[J]. 图书情报工作, 2013 , 57(04) : 120 -124,137 . DOI: 10.7536/j.issn.0252-3116.2013.04.021

Abstract

In the traditional citation analysis, it assumes that the relationship between literatures is simply based on citation links. For example, if paper A cites paper B, it means that B is important for A, while it is difficult to judge the contribution and citing reason. This paper treats with the citing paper, cited paper and citation context as the probability distribution on the topics, and analyzes the citing reason and contribution by full-text extraction. Firstly, it introduces the research background and significance, as well as the basic conceptions. Secondly, some methods are illustrated for supporting the experiment, such as full-text extraction, and topic model based on Labeled-LDA. Finally, different citation graphs are established and visualized based on different topics.

Key words： Labeled-LDA; topic model; citation analysis; full-text extraction

参考文献

[1] Gross P L K, Gross E. College libraries and chemical education [J]. Science, 1927,66(1713): 385-389.



[2] Garfield E.Citation analysis as a tool in journal evaluation[J]. Science, 1972,178(4060):471-479.



[3] Garfield E, Sher I H.New factors in the evaluation of scientific literature through citation indexing [J]. JASIST, 1963,14(3):195-201.



[4] Ma Nan, Guan Jiancheng, Zhao Yi. Bring PageRank to the citation analysis [J]. Information Processing and Management, 2008,44(2):800-810.



[5] Page L, Brin S, Motwani R, et al. The PageRank citation ranking: Bringing order to the Web [M].Stanford, California: Stanford InfoLab, 1998:1-17.



[6] Liu Xiaozhong, Zhang Jinsong, Guo Chun. Full-text citation analysis: A new method to enhance scholarly network[EB/OL].[2012-11-01].http://discern.uits.iu.edu:8790/publication/Full%20text%20citation.pdf.



[7] 曹俊, 晓云,廖顺宝. 基于正则表达式批量提取CNKI文献元数据技术探究[J].图书情报工作, 2010, 54(19):111-114.



[8] Blei D M, Ng A Y, Jordan M I. Latent dirichlet allocation [J]. The Journal of Machine Learning Research, 2003(3): 993-1022.



[9] Ramage D,Hall D,Nallapati R,et al. Labeled LDA: A supervised topic model for credit attribution in multi-labeled corpora[C]//Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing.Stroudsburg:Association for Computational Linguistics,2009:248-256.



[10] 李文波, 孙乐,张大鲲. 基于Labeled-LDA 模型的文本分类新算法[J].计算机学报, 2008, 31(4):620-627.



[11] Liu Xiaozhong, Zhang Jinsong, Guo Chun. Full-Text citation analysis: Enhancing bibliometrics and scientific publication ranking[C]//Proceedings of 21st ACM International Conference on Information and Knowledge Management. New York:ACM, 2012:1975-1979.

Options
文章导航

/

〈 〉