Microblog friends automatic clustering framework based on similarity measurement

In online social media like microblog, users can be easily overwhelmed by massive amount of information received from their friends. In this paper, we propose a framework to address this problem by recommending users clustering their friends into smaller groups, expecting messages from same groups are more similar than that from different groups. Firstly, profile, content and network structure features are used to capture the similarities of the friends respectively. Secondly, an unsupervised algorithm based on spectral clustering algorithm is employed to cluster the friends based on the similarity measurement. To improve the quality of clustering results, a clustering ensemble algorithm is adopted to combine all the clustering results obtained from these referred features. Experiments based on the data collected from Sina microblog are conducted to evaluate the accuracy and efficiency of the method. The results show that the proposed method can capture the friends' behavior characteristics efficiently and cluster them into proper groups.