Discovering patterns of advertisement propagation in Sina-Microblog

The explosive growth of microblogs has attracted many corporations and organizations. Microblogging has been considered as a high-quality advertising platform. In this study, we attempt to reveal the patterns of advertisement propagation in Sina-Microblog through analyzing a selected set of message cascades. Each message cascade is represented by a propagation tree and 33 features were extracted, which cover mainly three aspects of a cascade: the volume of the participants, the topology of the propagation paths, and the promptness of the propagation in term of time. To reveal the propagation patterns, We then group these message cascades using K-means clustering algorithm. Analysis of the resulted clusters reveals the patterns of advertisement propagation, based on which we further propose several metrics to measure the effectiveness of advertisement in microblogs.