期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2016
卷号：9
期号：4
页码：447-460
DOI：10.14257/ijhit.2016.9.4.39
出版社：SERSC
摘要：With the rapid development of Social Network, Micro-blog has becoming an important way for people communicating with each other and getting shared resource, so the research of Social Network especially Twitter and Micro-blog is crucial. The previous studies most focused on users' characteristic, topical issues, information broadcasting mechanism and so on. But there have been few studies to combine them together and find more valuable conclusions. In this paper, deep study of users' characteristic and their roles in topic diffusion is proposed. By using correlation analysis and K-Means clustering algorithm, ordinary users can be divided into several different types according to the data characteristics. Discriminate function has proved the rationality of classification in theory. By using PKUVIS, a visual analysis of the topic about the lots of Malaysia airline flight MH370 is applied to trace users' roles in topic diffusion. The empirical analysis show that, Micro-blog users assume different roles in the process of topic broadcasting mechanism based on individual attribute features.
