科技文档间非对称关系的双模态度量方法

doi:10.3969/j.issn.1000-1565.2021.05.016

摘要/Abstract

摘要： 针对科技文档间相似程度和影响力不同的问题,通过分析科技文档间非对称关系,提出一种新的度量方法.该方法将科技文档间非对称关系定义为文档覆盖度,并用公式覆盖度和文本覆盖度对其进行度量.公式覆盖度由改进的非对称因子计算,文本覆盖度通过利用文本的相对突出性调整特征向量的余弦夹角计算,公式覆盖度和文本覆盖度线性融合得到科技文档覆盖度.实验结果表明:与已有的2种科技文档关系度量方法相比,本文提出的非对称关系度量方法在聚类中的平均准确率分别提高了8%和4%.

关键词: 科技文档, 非对称性, 覆盖度, 关系度量

Abstract: Aiming at the problem of different degree of similarity and influence between scientific documents, a new measurement method is proposed by analyzing the asymmetric relationship between scientific documents.This method defines the asymmetric relationship between scientific documents as document coverage, and uses formula coverage and text coverage to measure it.The formula coverage is- DOI:10.3969/j.issn.1000-1565.2021.05.016科技文档间非对称关系的双模态度量方法徐建民,王鑫(河北大学网络空间安全与计算机学院,河北保定 071002)徐建民管理学博士,现任河北大学教授,河北大学计算机科学与技术专业硕士生导师,管理科学与工程专业、图书情报与档案专业博士生导师,入选河北省省管优秀专家,教育部科技委管理学部委员. 曾出版高等学校教材10部,学术专著1部,发表科研论文100余篇,获河北省教学成果一、二、三等奖各1项,国家教学成果奖1项,获河北省科技进步二、三等奖各1项,河北省社会科学优秀成果一等奖1项. 主要研究方向为信息检索、个性化推荐、在线社交网络分析、网络舆情监测与分析.摘要:针对科技文档间相似程度和影响力不同的问题,通过分析科技文档间非对称关系,提出一种新的度量方法.该方法将科技文档间非对称关系定义为文档覆盖度,并用公式覆盖度和文本覆盖度对其进行度量.公式覆盖度由改进的非对称因子计算,文本覆盖度通过利用文本的相对突出性调整特征向量的余弦夹角计算,公式覆盖度和文本覆盖度线性融合得到科技文档覆盖度.实验结果表明:与已有的2种科技文档关系度量方法相比,本文提出的非对称关系度量方法在聚类中的平均准确率分别提高了8%和4%.关键词:科技文档;非对称性;覆盖度;关系度量中图分类号:TP391 文献标志码:A 文章编号:1000-1565(2021)05-0587-12A double mode measurement method of asymmetric relationship between scientific documentsXU Jianmin, WANG Xin(College of Cyberspace Security and Computer, Hebei University, Baoding 071002, China)Abstract: Aiming at the problem of different degree of similarity and influence between scientific documents, a new measurement method is proposed by analyzing the asymmetric relationship between scientific documents.This method defines the asymmetric relationship between scientific documents as document coverage, and uses formula coverage and text coverage to measure it.The formula coverage is- 收稿日期:2021-03-05 基金项目:国家社会科学基金后期资助项目(17FTQ002);河北省自然科学基金资助项目(F2015201142) 第一作者:徐建民(1966—),男,河北邯郸人,河北大学教授,博士生导师,主要从事信息检索、在线社交网络方向研究.E-mail: hbuxjm@hbu.edu.cn 通信作者:王鑫(1996—),男,河北保定人,河北大学在读硕士研究生,主要从事信息检索方向研究.E-mail:838148632@qq.com第5期徐建民等:科技文档间非对称关系的双模态度量方法calculated by an improved asymmetric factor, and the text coverage is calculated by using the relative prominence of the text to adjust the cosine angle of the feature vector, The formula coverage and text coverage are linearly fused to obtain the coverage of scientific documents.The experimental results show that compared with the existing two measurement methods for relationship between scientific documents, the average accuracy of the asymmetric relationship measurement method proposed in this paper is improved by about 8 percentage points and 4 percentage points.

Key words: scientific documents, asymmetry, coverage, relational metric

中图分类号:

TP391

徐建民,王鑫. 科技文档间非对称关系的双模态度量方法[J]. 河北大学学报(自然科学版), 2021, 41(5): 587-598.

XU Jianmin, WANG Xin. A double mode measurement method of asymmetric relationship between scientific documents[J]. Journal of Hebei University(Natural Science Edition), 2021, 41(5): 587-598.

参考文献

[1] 陈二静,姜恩波.文本相似度计算方法研究综述[J].数据分析与知识发现,2017,1(6):1-11.DOI:10.11925/infotech.2096-3467.2017.06.01.
[2] GARG A, ENRIGHT C G, MADDEN M G. On asymmetric similarity search[C] // IEEE International Conference on Machine Learning & Applications, IEEE, 2015.
[3] 徐建民,许彩云.基于文本和公式的科技文档相似度计算[J].数据分析与知识发现,2018,2(10):103-109.DOI:10.11925/infotech.2096-3467.2018.0211.
[4] RORISSA A. Relationships between perceived features and similarity of images: A test of Tversky's contrast model[J]. Journal of the American Society for Information ence and Technology, 2007, 58(10):1401-1418.DOI: 10.1002/ais.20606.
[5] 刘勘,周丽红,陈譞.基于关键词的科技文献聚类研究[J].图书情报工作,2012,56(4):6-11.
[6] 牛奉高,李星.共现潜在语义向量空间模型的进一步研究[J].情报杂志,2017,36(12):166-172.DOI:10.3969/j.issn.1002-1965.2017.12.027.
[7] JIANG W, SAMANTHULA B K. N-gram based secure similar document detection[C] // Data & Applications Security & Privacy Xxv- Ifip Wg 113 Conference, Springer-Verlag, 2011.
[8] 楼雯,王慧,鞠源.基于二值相似度计算的异构本体融合方法[J].情报学报,2019,38(06):622-631.DOI:10.3772/j.issn.1000-0135.2019.06.007.
[9] 黎雪微,应时,洪伟.基于本体和信息量融合的个性化推荐方法研究[J].情报科学,2019,37(9):90-95.
[10] 朱戈.一种基于PageRank的文献相似性搜索算法[J].计算机工程与应用,2013,49(8):142-145.DOI:10.3778/j.issn.1002-8331.1108-0414.
[11] 翟玲,崔旭.基于分段估计和PageRank的文本信息相似性搜索算法研究[J/OL].[2021-05-09].吉林大学学报(工学版):1-6.https://kns.cnki.net/kcms/detail/22.1341.T.20210320.1549.002.html.
[12] CHEN H, LYNCH K J. Automatic construction of networks of concepts characterizing document databases[J]. IEEE Transactions on Systems, Man and Cybernetics, 1992, 22(5):885-902.
[13] GUAN R, SHI X, MARCHESE M, et al. Text clustering with seeds affinity propagation[J]. IEEE Transactions on Knowledge and Data Engineering, 2011,23(4):627-637.
[14] CHUA W W K, GOH A E S. An asymmetric similarity measure for ontologies based on the feature contrast model[C] // International Conference on Complex, IEEE, 2010.
[15] ALBERTONI R, MARTINO M D. Asymmetric and context-dependent semantic similarity among ontology instances[J]. lecture notes in computer science, 2008,4900(10):1-30.
[16] YOSHIDA H, SHIDA T, KINDO T. Asymmetric similarity with modified overlap coefficient among documents[C] // IEEE Pacific Rim Conference on Communications, Computers & Signal Processing, IEEE, 2001.DOI:10.1109/PACRIM.2001.953532.
[17] TEITELBAUM, JOSHUA C. Asymmetric empirical similarity[J]. Mathematical Social Sciences,2013, 66(3):346-351.
[18] 宋韶旭, 李春平. 基于非对称相似度的文本聚类方法[J]. 清华大学学报(自然科学版),2006(7):1325-1328.DOI:10.3321/j.issn:1000-0054.2006.07.037.
[19] WU Y, WANG S, SONG G, et al. Online asymmetric metric learning with multi-layer similarity aggregation for cross-modal retrieval[C] //IEEE Transactions on Image Processing, 2019:1-1.DOI:10.1109/TIP.2019.2908774.
[20] 陈凯燕.中文文档图像的数学公式定位[D].武汉:华中师范大学,2016.
[21] 许彩云.基于文本和公式的科技文档相似度计算研究[D].保定:河北大学,2019.
[22] MOUSSA M, MANDOIU, ION I. Single cell RNA-seq data clustering using TF-IDF based methods[J]. Bmc Genomics, 2018, 19(S6):127-141.
[23] TVERSKY. A features of similarity[J]. Readings in Cognitive Science, 1988, 84(4):290-302.
[24] 唐亚伟.公式相似度算法及其在论文查重中的应用研究[D].锦州:渤海大学,2013.
[25] JOHANNESSON M. Modelling asymmetric similarity with prominence[J]. British Journal of Mathematical and Statistical Psychology, 2000, 53(1):121-139.
[26] 宰新宇,田学东.基于公式描述结构和词嵌入的科技文档检索方法[J].数据分析与知识发现,2020,4(1):131-138.
[27] AHMADZADEHGOLI N, MOHAMMADPOUR A, BEHZADI M H. LINEX K-Means: Clustering by an Asymmetric Dissimilarity Measure[J]. Journal of statistical Theory and Applications, 2018,17(1):29.DOI:10.2991/jsta.2018.17.1.3.
[28] 徐建民,王平.小型中文信息检索测试集的构建与分析[J].情报杂志,2009,28(1):13-16.DOI:10.3969/j.issn.1002-1965.2009.01.004.
[29] 韩忠明,陈妮, 张慧, 等. 一种非对称距离下的层次聚类算法[J].模式识别与人工智能,2014,27(5):410-416.DOI:10.3969/j.issn.1003-6059.2014.05.004. (