融入主题和时间因素的文献影响力评价研究
|
摘要
【目的/意义】本文综合考虑文献的主题、被引频次和发文时间,从三维视角出发,提出一种文献影响力指数(Influence Index of Document,IIOD)计算方法,在主题分类的基础上对文献的影响力进行评价。【方法/过程】以中国引文数据库中2013-2017年情报学领域9本期刊的10523篇文献为样本,利用LDA模型进行主题提取,并将LDA与K-means算法结合实现文本分类,进而计算特定主题下文献的影响力指数,根据文献影响力指数对文献进行排名。【结果/结论】文献影响力指数综合考虑了文献的主题特征、被引频次和时效性,其排名结果更加全面合理,为评价文献的影响力提供了一种新的视角。
【Purpose/significance】From a three dimensional perspective, a literature influence index calculation method is proposed, which is based on the theme, cited frequency and time of writing, and evaluates the influence of the literature on the basis of the subject classification.【Method/process】Taking 10523 articles of 9 information science periodicals in the Chinese citation database from 2013-2017 as samples, the LDA model is used to extract the theme, the text classification is realized by combining the LDA and K-means algorithm,the influence index of the literature under the specific subject is calculated, and then the literature is ranked according to the literature impact force index.【Result/conclusion】The literature influence index comprehensively considers the theme, cited frequency and timeliness of the literature, and its results are more comprehensive and reasonable, which provides a new perspective for evaluating the influence of literature.
【Purpose/significance】From a three dimensional perspective, a literature influence index calculation method is proposed, which is based on the theme, cited frequency and time of writing, and evaluates the influence of the literature on the basis of the subject classification.【Method/process】Taking 10523 articles of 9 information science periodicals in the Chinese citation database from 2013-2017 as samples, the LDA model is used to extract the theme, the text classification is realized by combining the LDA and K-means algorithm,the influence index of the literature under the specific subject is calculated, and then the literature is ranked according to the literature impact force index.【Result/conclusion】The literature influence index comprehensively considers the theme, cited frequency and timeliness of the literature, and its results are more comprehensive and reasonable, which provides a new perspective for evaluating the influence of literature.
引文
1 Kam C. Chan,Pikki Lai,Kartono Liano. A threshold citation analysis in marketing research[J]. European Journal of Marketing,2012,46(1):134-156.
2 许琦.一种基于引证网络的文献影响因子计算方法[J],情报理论与实践,2011,34(7):98-102.
3 Hirsch J E. An index to quantify an individual’s research output[J].Proceedings of National Academy of Sciences,2005,102(46):16569-16572.
4 Mikalai Krapivin,Maurizio Marchese,Fabio Casati.Complex Sciences[M].Berlin:Springer Berlin Heidelberg:2009.
5 曾文,桂婕,徐红姣,刘敏.基于领域的科技文献重要度评估方法研究[J].情报理论与实践,2015,(12):73-76,92.
6 贾宁.文献被引的年代分布对被引文献评价的意义——以物理学科为例[J].图书馆杂志,2016,(12):55-62.
7 汪志伟,邹艳妮,吴舒霞. PageRank算法应用在文献检索排序中的研究及改进[J].情报理论与实践,2016,(11):126-130,144.
8 翟晓芳,刘全明,程耀东,李海波.结合社会化网络的文献综合计量模型[J].计算机工程,2016,(6):21-26.
9 马秀峰,郭顺利,宋凯.基于LDA主题模型的“内容-方法”共现分析研究——以情报学领域为例[J].情报科学,2018,36(4):69-74.
10 李湘东,张娇,袁满.基于LDA模型的科技期刊主题演化研究[J].情报杂志,2014,33(7):115-121.
11 曹娟,张勇东,李锦涛,唐胜.一种基于密度的自适应最优LDA模型选择方法[J].计算机学报,2008,(10):1780-1787.
12 宋凯,李秀霞,赵思喆.基于CTM模型与K-means算法融合的文本聚类研究[J].情报理论与实践,2017,40(11):135-138.
2 许琦.一种基于引证网络的文献影响因子计算方法[J],情报理论与实践,2011,34(7):98-102.
3 Hirsch J E. An index to quantify an individual’s research output[J].Proceedings of National Academy of Sciences,2005,102(46):16569-16572.
4 Mikalai Krapivin,Maurizio Marchese,Fabio Casati.Complex Sciences[M].Berlin:Springer Berlin Heidelberg:2009.
5 曾文,桂婕,徐红姣,刘敏.基于领域的科技文献重要度评估方法研究[J].情报理论与实践,2015,(12):73-76,92.
6 贾宁.文献被引的年代分布对被引文献评价的意义——以物理学科为例[J].图书馆杂志,2016,(12):55-62.
7 汪志伟,邹艳妮,吴舒霞. PageRank算法应用在文献检索排序中的研究及改进[J].情报理论与实践,2016,(11):126-130,144.
8 翟晓芳,刘全明,程耀东,李海波.结合社会化网络的文献综合计量模型[J].计算机工程,2016,(6):21-26.
9 马秀峰,郭顺利,宋凯.基于LDA主题模型的“内容-方法”共现分析研究——以情报学领域为例[J].情报科学,2018,36(4):69-74.
10 李湘东,张娇,袁满.基于LDA模型的科技期刊主题演化研究[J].情报杂志,2014,33(7):115-121.
11 曹娟,张勇东,李锦涛,唐胜.一种基于密度的自适应最优LDA模型选择方法[J].计算机学报,2008,(10):1780-1787.
12 宋凯,李秀霞,赵思喆.基于CTM模型与K-means算法融合的文本聚类研究[J].情报理论与实践,2017,40(11):135-138.