文章摘要

刘自强,王效岳,白如江.多维度视角下学科主题演化可视化分析方法研究——以我国图书情报领域大数据研究为例[J].中国图书馆学报,2016,42(6):67~84
多维度视角下学科主题演化可视化分析方法研究——以我国图书情报领域大数据研究为例
Research on Visualization Analysis Method of Discipline Topics Evolution from the Perspective of Multi Dimensions:A Case Study of the Big Data in the Field of Library and Information Science in ChinaL
投稿时间:2016-06-03  修订日期:2016-07-29
DOI:
中文关键词: 进行语义角色分类,利用Fast Unfolding算法识别出具有语义特征的学科主题  利用余弦相似度计算公式计算学科主题相似度判定演化关系  构建多维度学科主题演化分析模型,并设计了三种创新性的科学知识图谱,进行学科主题强度、结构和内容三个维度的可视化分析,通过相互作用可以帮助快速消化、理解信息和精炼分析结果,有效地分析学科主题演化的复杂过程。通过对我国图书情报领域近10年大数据研究的实证分析,证明该方法具有可行性和有效性。图9。表4。参考文献37。关键词  学科主题演化  语义角色标注  社区发现算法  可视化
英文关键词: Discipline topics evolution  Semantic role labeling  Community discovery algorithm  Visualization
基金项目:本文系国家社会科学基金项目“未来新兴科学研究前沿识别研究”(编号:16BTQ083)的研究成果之一
作者单位
刘自强 山东理工大学科技信息研究所 山东 淄博 255049 
王效岳 山东理工大学科技信息研究所 山东 淄博 255049 
白如江 山东理工大学科技信息研究所 山东 淄博 255049 
摘要点击次数: 1817
全文下载次数: 299
中文摘要:
      探测、识别某学科领域研究主题的演化过程并进行可视化分析,对于掌握研究现状和发展趋势具有重要意义。学科主题演化是一个复杂过程,存在多种变量,如主题强度、结构和内容等,目前研究主要以单一维度进行可视化分析,信息负荷过大,存在感知局限性。本文提出多维度视角下学科主题演化可视化分析方法:通过人工标注方法对关键词进行语义角色分类,利用Fast Unfolding算法识别出具有语义特征的学科主题;利用余弦相似度计算公式计算学科主题相似度判定演化关系;构建多维度学科主题演化分析模型,并设计了三种创新性的科学知识图谱,进行学科主题强度、结构和内容三个维度的可视化分析,通过相互作用可以帮助快速消化、理解信息和精炼分析结果,有效地分析学科主题演化的复杂过程。通过对我国图书情报领域近10年大数据研究的实证分析,证明该方法具有可行性和有效性。图9。表4。参考文献37。
英文摘要:

    Detection and identification of the evolution of research topics in a discipline has important significance for researchers to grasp its research status and development trend. Visual analysis can show the relationship between themes based on topics recognition,help users to enhance their perception and cognition,and to find useful information quickly in a field on the research status,research hotspots and development trends,and to digest,understand and effectively analyze vast amounts of information. However,discipline topics evolution is a complex process and there are many variables,such as the intensity,structure and content of topics. The single dimension visualization analysis causes the information overload,leading to three problems:perceptive limitations,cognitive limitations,and performance limitations.

    This paper presents a visualization analysis method of discipline topics evolution from a multidimensional perspective:using the artificial annotation method to make semantic role classification of keywords,using Fast Unfolding algorithm recognition with the semantic features to identify the topics;using cosine similarity to calculate the formula of similarity between topics evolution;constructing evolution analysis model of multidimensional discipline topics,and designing three innovative scientific knowledge map by using JavaScript and Web front-end visualization technology to analyze the visualization of the intensity,structure and content of the topics. Through the interaction of the core areas of research questions,the evolution path and trend of research methods and key technologies of the topics,and the macro evolution trend,meso evolution process and microscopic evolution details can be revealed. It can effectively help to quickly digest and understand information and refine the topic evolution analysis results to reveal the complex process of topics evolution.

    The experiment on ‘big data study in library and intelligence field in the past 10 years’ proves the visualization analysis method proposed in this paper could effectively demonstrate the complicated process of topics evolution in a discipline. Compared with the other visualization analysis methods,the proposed method based on topic strength,structure and internal basic knowledge unit evolution could visualize and analyze the core research points,the primary research methods,the key technology topics evolution path and trends and the evolution trend from the macro-angle,the evolution process from meso-angle and the evolution particulars from the micro angle,so as to better analyze the complicated topics evolution process. The proposed method is applicable only to scientific papers including key words and needs to be expanded to include other data resources.

    The practical significance lies in:1)Proposing the dynamic topics identification method to provide reference to related research;2)Proposing a multidimensional topics evolution analysis model and an innovative visualization method which can be used to analyze the topics evolution rules and discover scientific and technological knowledge,etc;3)The findings can be used in scientific and research management to support decision-making,raise research efficiency and help to promote the scientific and technological innovation in related fields. 9 figs. 4 tabs. 37 refs.

查看全文   查看/发表评论  下载PDF阅读器