| 基于科技论文多特征项共现突发强度分析方法的算法实现与可视化图谱研究 |
| |
| 引用本文： | 庞弘燊.基于科技论文多特征项共现突发强度分析方法的算法实现与可视化图谱研究[J].图书情报工作,2015,59(24):115-122. |
| |
| 作者姓名： | 庞弘燊 |
| |
| 作者单位： | ISTIC-THOMSON REUTERS科学计量学联合实验室 北京 100038;中国科学院广州生物医药与健康研究院 广州 510530 |
| |
| 基金项目： | 本文系ISTIC-THOMSON REUTERS科学计量学联合实验室开放基金项目"基于学科领域多特征项共现的知识图谱分析方法设计与实现——以WoS学科数据为例"和国家自然科学基金管理学部青年项目"基于学科领域科技论文多重共现的情报计量分析方法研究"(项目编号:71403261)研究成果之一。 |
| |
| 摘 要： | 目的/意义]基于科技论文多特征项共现突发强度的分析方法研究是将各学科领域科技论文文献载体中的多特征项共现信息定量化、重点热点突发的信息内容可视化的知识图谱分析方法。从动态论文等文献的文档流中探测出突发的特征项对识别密集的内容、活跃的特征项以及预测文本内容的发展走势具有重要的意义。方法/过程]本研究针对科技论文多特征项共现的突发监测问题,对比目前已有的突发监测分析算法,将改进后的基于卡方统计的热点词计算方法进一步应用于本研究所设计的多特征项突发共现分析方法,并自主开发多特征项突发共现可视化分析工具,用于科技论文多特征项突发共现的图谱可视化,以期通过该研究来揭示相关科技文献的变化状况及突发的热点内容。结果/结论]通过将本方法应用到科研机构年度发表论文的监测当中,可以监测分析科研机构发文作者、关键词、发表期刊及其相互间关系(如作者-关键词等)在各年的突发情况,并能通过该分析方法以及交叉图谱进一步解读突发特征项的含义,并能揭示出比分析单一特征项突发情况更为广泛和深入的知识内容。
|
| 关 键 词： | 多特征项共现 突发强度分析 科技论文监测 |
| 收稿时间： | 2015-11-06 |
Research on the Algorithms and Visualization Analysis Tool of the Burst Detection Analysis Method Based on Multiple Relationships in Collections of Journal Papers |
| |
| Pang Hongshen.Research on the Algorithms and Visualization Analysis Tool of the Burst Detection Analysis Method Based on Multiple Relationships in Collections of Journal Papers[J].Library and Information Service,2015,59(24):115-122. |
| |
| Authors: | Pang Hongshen |
| |
| Institution: | ISTIC-THOMSON REUTERS Joint Lab for Scientometrics Research, Beijing 100038;Guangzhou Institutes of Biomedicine and Health, Chinese Academy of Sciences, Guangzhou 510530 |
| |
| Abstract: | Purpose/significance] The burst detection analysis method based on multiple relationships in collections of journal papers is a knowledge discovery method for multiple co-occurrences and hot topics in scientific and technicalpapers. It is significant to detect the research content and predict the research trends in papers' data stream.Method/process] This paper focused on the burst detection problem of multiple co-occurrences in collections of journal papers. By comparing the algorithms of burst detection, the author selected a suitable algorithm in this research and developed the visualization analysis tool and designs the analysis model for burst detection analysis on multiple co-occurrences. This paper further applied the multiple co-occurrences burst detection method to reveal the research trends and hot topics in a set of institutes' papers.Result/conclusion] By applying this method to the annual published papers' detection of scientific research institutions, we can analyze the burst detection situation of authors, key words, published journals and their relationships (such as the authors-keywords, etc.). We can also dig out deeper information contents than the general burst detection analysis method. |
| |
| Keywords: | multiple co-occurrences burst detection analysis journal papers analysis |
|
| 点击此处可从《图书情报工作》浏览原始摘要信息 |
| 点击此处可从《图书情报工作》下载免费的PDF全文 |
