| 基于MUK-means算法的微博舆情意见领袖群识别 |
| |
| 引用本文： | 李熠辉,李 冠,赵卫东.基于MUK-means算法的微博舆情意见领袖群识别[J].教育技术导刊,2019,18(12):30-34. |
| |
| 作者姓名： | 李熠辉 李 冠 赵卫东 |
| |
| 作者单位： | 山东科技大学 计算机科学与工程学院，山东 青岛 266590 |
| |
| 基金项目： | 山东省研究生教育创新计划一般项目（SDYC16022） |
| |
| 摘 要： | 推动微博舆情事件演化是众多意见领袖共同作用的结果，因此识别意见领袖群对于舆情事件的监管具有重要作用。提出微博舆情话题下的意见领袖群识别模型，综合考虑用户属性特征、交互特征和网络结构，设计微博舆情下用户影响力评估算法MUR，并结合K-means算法形成MUK-means算法，实现对意见领袖群的识别。以新浪微博数据进行实验，MUK-means算法的聚类时间（14s）远远少于传统K-means算法（32s），而且基于MUK-means算法得到的意见领袖群的用户覆盖率高达86.3%。实验结果表明，MUK-means算法改进了K-means算法初始聚类中心不确定的缺点，不仅提高了聚类效率，而且实现了对意见领袖群的有效识别。
|
| 关 键 词： | 微博舆情 MUR MUK-means 意见领袖群 |
| 收稿时间： | 2019-07-22 |
Research on Weibo Public Opinion Leader Group Recognition Based on MUK-means Algorithm |
| |
| LI Yi-hui,LI Guan,ZHAO Wei-dong.Research on Weibo Public Opinion Leader Group Recognition Based on MUK-means Algorithm[J].Introduction of Educational Technology,2019,18(12):30-34. |
| |
| Authors: | LI Yi-hui LI Guan ZHAO Wei-dong |
| |
| Institution: | College of Computer Science and Engineering，Shandong University of Science and Technology，Qingdao 266590，China |
| |
| Abstract: | Promoting the evolution of public opinion events is the result of the interaction of many opinion leaders. So the identification of opinion leader group plays an important role in the supervision of public opinion events. In this paper， the opinion leader group recognition model was put forward under the micro-blogging public opinion. By considering the user attribute characteristics， the interaction characteristics and the network structure， the Microblog-lyric User-Influence Rank（MUR） algorithm was designed. Then MUR algorithm combined with the K-means algorithm to form the MUK-means algorithm， which realized the recognition of the opinion leader group. Experiments with Sina Weibo data show that the clustering time of the MUK-means algorithm （14s） is far less than the traditional K-means algorithm （32s）， and the user coverage rate of the opinion leader group based on the MUK-means algorithm is as high as 86.3%. The MUK-means algorithm improves the uncertainty of the initial clustering center of K-means algorithm， which not only improves the clustering efficiency， but also realizes the effective recognition of the opinion leader group. |
| |
| Keywords: | Micro-blogging public opinion MUR MUK-means opinion leader group |
|
| 点击此处可从《教育技术导刊》浏览原始摘要信息 |
| 点击此处可从《教育技术导刊》下载免费的PDF全文 |
