基于动态复杂网络分析方法的事件检测实证研究

Empirical Research on Event Detection Based on Dynamic Complex Network Analysis Method

[目的/意义]人们生活在复杂多变的动态社会系统中,产生了大规模社会关系数据,从中进行情报分析具有重要的研究意义与应用价值,同时也面临着巨大挑战。依托于动态复杂网络中社区演化建模方法,能够挖掘动态社区及其演化模式,进一步发现社区演化异常,可应用于真实社会网络中的事件检测,成为当前情报分析的重要手段。[方法/过程]针对社会复杂系统中的两个真实场景,基于真实动态社会复杂网络,应用前沿动态社区演化建模方法,在未知事件信息的情况下挖掘社区演化模式并计算演化强度,进而发现演化异常,实现社会复杂系统中的事件检测。[结果/结论]实证研究充分表明:社会网络结构演化突变的驱动力来源于背后的真实事件。同时,也验证了动态社区演化模型对于社会复杂系统中事件检测的有效性。

[Purpose/Significance]People live in a complex and changeable dynamic social system,which produces large-scale social relationship data.It is of great research significance and application value to carry out intelligence analysis from it,and it also faces great challenges.Relying on the community evolution modeling method in the dynamic complex network,it can mine the dynamic community and its evolution pattern,and further discover the community evolution anomaly.It can be applied to the event detection in the real social network,and becomes an important means of current intelligence analysis.[Method/Process]Aiming at two real scenes in the complex social system and based on the real dynamic social complex network,this paper applies the cutting-edge dynamic community evolution modeling method to mine the community evolution pattern and calculate the evolution intensity under the circumstance of unknown event information,and then discover the evolution anomaly,and realize the event detection in the complex social system.[Result/Conclusion]The empirical research fully shows that the driving force of social network structure evolution mutation comes from the real events behind.At the same time,the validity of the dynamic community evolution model for event detection in complex social systems is verified.