结合时间信息的作者耦合分析方法

Towards the Improvements of Author Bibliographic Coupling Analysis Combined with Time Information

[目的/意义]作者耦合分析是绘制科学知识图谱和了解学科结构一种重要方法,主要使用了作者间共同引用相同参考文献的篇数来进行定义。传统作者耦合分析算法较为多样,包括简单计算法、最小值计算法和组合计算法等,其中以最小值计算法最为精确。然而,尽管传统作者耦合分析能够探测出学科前沿,但因为原始矩阵输入信息量较小,其绘制出的图谱包含的内容和效度都较为有限。该文类比作者共引分析中的改进研究,尝试在基于最小值计算法的基础上融入原文和引文的发表时间信息,对传统作者耦合分析方法进行改进。[方法/过程]使用了包括网络分析在内的实证研究,绘制了不同方法下的科学知识图谱,并对结果进行了评估。[结果/结论]结果显示,融入原文和引文发表时间信息的作者耦合分析方法能展示出更多细节信息,图谱的聚类效果也较传统方法更优。

[Purpose/Significance] Author bibliographic coupling analysis （ABCA） is an important approach in mapping knowledge do- mains and depicting scientific intellectual structures. It is defined as the number of references co-listed by two authors＂ publications. Tradi- tional ABCA has various calculation algorithms, such as simple method, minimum method, and combined method, in which the minimum method shows the most accurate result. However,although ABCA is used to explore the frontier of disciplines,the content and validity of the knowledge domain maps could be limited because of the uninformative input of its raw matrix. By analogy with previous work,this pa- per tries to combine published time information of citing and cited papers to improve traditional ABCA method. [ Method/Process] We do an empirical study including network analysis and test the results by showing knowledge domain mappings via different methods. [ Result/ Conclusion ] The results show that our newly proposed method can display more details and the maps drawn have a better ability in cluste- ring.