一种关于交叉学科领域话题挖掘的可视分析方法

发布时间：2018-02-03 08:52

本文关键词： 话题挖掘文本可视化可视分析　出处：《天津大学》2016年硕士论文　论文类型：学位论文

【摘要】：交叉科研领域的话题挖掘可以帮助用户发现相关研究领域之间的关系,同时让用户得到对相关研究领域的概览。有些研究领域具有与其它研究领域相类似的研究话题,所以一个研究领域内部发表的文章成果或发现也可能对另一个研究领域的研究产生影响。交叉学科领域的话题演变可以帮助研究者们从不同的角度来重新认识所研究的学科,但是在之前的工作中很少有人关注交叉学科领域话题演变的研究。本文针对交叉科研领域的话题演化,设计了一个可视分析系统,并综合概率主题模型对交叉科研领域的文本数据进行了研究。本文采用了一个层次型的话题模型来抽取不同学科领域的研究话题,并且利用数学模型将这些话题构造关联关系,发现了研究领域的研究话题的变化和趋势。另外,本文设计了一个简单高效的可视化界面,使用了包括词云、Sankey图、Treemap以及散点图等可交互和联动的视图,并利用其提供的交互操作来帮助用户更深层次地理解领域的发展趋势以及三个领域之间的关系。本文还以可视化,图形学和数据挖掘三个学科为例,设计了新话题的发现、交叉领域演化分析、交叉领域话题分析三个案例来进行研究和分析。本文研究了跨学科研究领域的话题演化关系,并且使用可视分析方法来判定属于各学科领域内部的和交叉不同学科领域的研究话题。用户利用可视分析系统,通过一系列视图与视图之间的联动交互,利用几个视图显示的结果探索和判断一些特定的话题,从而得到关于交叉科研领域的话题演化的结论。三个案例研究的结果证明了本文提出的方法的有效性。
[Abstract]:Cross-research topic mining can help users to discover the relationship between related research areas. At the same time, give users an overview of the relevant areas of research. Some research areas have similar research topics with other research areas. So the results or findings of articles published within one field of study may also have an impact on research in another field. The evolution of cross-disciplinary topics can help researchers to re-examine what they are doing from different perspectives. The subject of research. However, in the previous work, few people pay attention to the topic evolution of cross-disciplinary domain. In this paper, a visual analysis system is designed for the topic evolution in cross-research field. A hierarchical topic model is used to extract the research topics from different disciplines. In addition, this paper designs a simple and efficient visual interface, which includes the word cloud. Sankey diagrams can be interactive and interactive views such as Treemap and scatter diagrams. The interaction provided by it can help users to understand the trend of the development of the field and the relationship between the three fields at a deeper level. This paper also takes the visualization, graphics and data mining as an example. This paper designs three cases to study and analyze the discovery of new topics, cross-domain evolution analysis and cross-domain topic analysis. And the visual analysis method is used to determine the research topics that belong to each subject domain and cross different subject fields. Users use visual analysis system through a series of views and view interaction. The results of several views are used to explore and judge some specific topics, and the results of three case studies prove the validity of the proposed method.
【学位授予单位】：天津大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.1

【相似文献】