Partitioning sparse matrices with eigenvectors of graphs
SIAM Journal on Matrix Analysis and Applications
Learning probabilistic models of link structure
The Journal of Machine Learning Research
Structure and evolution of online social networks
Proceedings of the 12th ACM SIGKDD international conference on Knowledge discovery and data mining
The link-prediction problem for social networks
Journal of the American Society for Information Science and Technology
Unsupervised prediction of citation influences
Proceedings of the 24th international conference on Machine learning
A measurement-driven analysis of information propagation in the flickr social network
Proceedings of the 18th international conference on World wide web
Combining link and content for community detection: a discriminative approach
Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining
What is Twitter, a social network or a news media?
Proceedings of the 19th international conference on World wide web
Hi-index | 0.00 |
In this paper, we focus on the problem of community detection on TSina: the most popular microblogging network in China. By characterizing the structure and content of microgroup (community) on TSina in detail, we reveal that different from ordinary social networks, the degree assortativity coefficients are negative on most microgroups. In addition, we find that users from the same microgroup likely exhibit some similar attributes (e.g., sharing many followers, tags and topics). Inspired by these new findings, we propose a united method for microgroup detection without losing the information of link structure and user attribute. First, the link direction is converted to the weight by giving higher value to the more surprising link, while attribute similarity between two users is measured by the Jaccard coefficient of common features like followers, tags, and topics. Then, above two factors are uniformly converted to the edge weight of a newly generated network. Finally, many frequently used community detection algorithms that support weighted network would be employed. Extensive experiments on real social networks show that the factors of link structure and user attribute play almost equally important roles in microgroup detection on TSina. Our newly proposed method significantly outperforms the traditional methods with average accuracy being improved by 25%, and the number of unrecognized users decreasing by about 75%.