基于时间序列相似性的网络社区检测方法

Community detection method of networks based on time series similarity

社区检测作为目前复杂网络的研究热点之一,其检测结果能帮助人们深入理解复杂网络的网络结构和内在运行机制,并具有非常高的应用价值。随着数据采集等技术的不断发展,复杂系统中的个体所具有的海量时间序列数据得以保存。本文针对一些具有时间序列数据的复杂系统,提出根据时间序列之间的相似性重构出其对应的复杂网络,并利用阈值法将网络进行了相应的简化,最后利用社区检测算法将网络划分为不同的社区,从而对复杂网络的网络拓扑结构和社区结构进行理解和分析。利用上证180指数成分股票的收盘价时间序列数据对该方法进行了实验分析验证,结果表明了该方法能够有效地检测出网络中的社区结构。

As one of the current research hotspots in complex networks,community detection results can help people deeply understand the network structure and internal operation mechanism of complex networks,and have very high application value.With the continuous development of data collection and other technologies,the massive time series data possessed by individuals in complex systems can be preserved.Based on the above background,this article proposes to construct a network corresponding to a complex system using the similarity between individual time series data and simplifies the network by using the threshold method.Finally,the network is divided into different communities by using the community detection algorithm,so as to understand and analyze the network topology and community structure of the complex network.At the same time,this article conducted experimental analysis and validation on this method using the closing price time series data of the Shanghai 180 Index component stocks,and the results showed that this method can effectively detect community structures in the network.