Exploring the efficiency and mechanism of whistleblowing system on social networking site

Malicious information like harassment, slander or bogus news is becoming a serious problem on social networking sites, and solving it requires profound insight into user behaviors of malicious and victim accounts. In this paper, we conduct a preliminary study on a whistleblowing network, constructed from 328472 whistleblowing (complaining) reports. To evaluate its efficiency, we perform a series of network analysis. Macroscopically, it shows weak connectivity, strong power law degree distribution; microscopically, it exhibits interesting micro structural patterns. The indiscriminate treatment of whistleblowees with low and high degree indicates that there is still much room to increase the efficiency of the current system. To understand the link between the macro and micro characteristics and the network forming mechanism, we propose a generating algorithm framework called Dynamic Configuration Framework. The simulation result demonstrates that the micro structural patterns are naturally formed along with the macro characteristics and the degree distribution can be qualitatively but not quantitatively explained by linear preferential attachment. The capability of our model to predict dynamic network evolution is the key to implement a new proactive defending layer. These findings provide a deep understanding of the whistleblowing network, and can give guidelines to improve the system fighting against malicious information.