| 连边社团检测算法对共词分析聚类结果的改进研究 |
| |
| 引用本文： | 孙海生.连边社团检测算法对共词分析聚类结果的改进研究[J].图书情报工作,2016,60(10):123-129. |
| |
| 作者姓名： | 孙海生 |
| |
| 作者单位： | 聊城大学图书馆 聊城 252059 |
| |
| 摘 要： | 目的/意义] 传统共词分析的聚类算法存在以下不足:①关键词只能被划归一个聚类;②聚类过程对分类数目的确定缺乏严格判断标准。针对以上问题,采用复杂网络理论进行改进研究。方法/过程] 采用连边社团检测算法对关键词进行聚类,以科学计量学为例进行实证研究。结果/结论] 分析结果表明:算法对关键词的聚类结果有较好的改进效果,能够把核心度高的关键词同时划分到不同的研究主题之中,克服传统聚类算法的不足,而且划分密度可为聚类数目的确定提供客观判断依据。
|
| 关 键 词： | 共词分析 复杂网络 连边社团检测算法 科学计量学 |
| 收稿时间： | 2016-01-27 |
Study on Improvement of Clustering Results of Co-Word Analysis Based on Link Communities Algorithm |
| |
| Sun Haisheng.Study on Improvement of Clustering Results of Co-Word Analysis Based on Link Communities Algorithm[J].Library and Information Service,2016,60(10):123-129. |
| |
| Authors: | Sun Haisheng |
| |
| Institution: | Library of Liaocheng University, Liaocheng 252059 |
| |
| Abstract: | Purpose/significance] There are some shortages as follows in the present study on co-word analysis:1) every keyword belonging to only one cluster; 2) lack of strict criteria for determining the number of clusters. Complex network theory is used to solve the above problems.Method/process] Link communities algorithm is applied to improve result of clustering, and an empirical research of scientometrics is conducted.Result/conclusion] The results show that the improvement effect on clustering is notable. Key words with high coreness simultaneously belong to several communities, which overcomes the shortages of traditional clustering algorithm. Moreover, partition density provides an objective basis for the determination of the number of clusters. |
| |
| Keywords: | co-word analysis complex network link communities algorithm scientometrics |
|
| 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 |
