Research On Dynamic Characteristics Of Mobile Communication Time Series Based On Complex Network Modeling

With the development of mobile communication technology,various applications and terminals begin to access the network on a large scale,and a large amount of mobile communication data is accumulated in the system.Mobile communication data includes throughput and voice traffic,which respectively represent the dynamic changes of data and voice generated by people's communication behavior in the geographical area covered by the communication base station.Research on mobile communication data based on complex network method is helpful to mine valuable information contained in base station data from a new perspective.Based on the data of mobile communication base station in a certain area,we mainly do the following research:(1)The similarity of mobile communication time series.According to the time series similarity measurement methods,we measure the similarity of throughput dynamic changes of base stations,and construct Pearson correlation network,Spearman correlation network and Mutual information network respectively.The discrete distribution of node degree,node betweenness and node clustering coefficient of the three networks is visualized.The network topology reflects that the non-linear relationship between traffic time series of base station is stronger than the linear relationship.In addition,the network topology reflects that the base station system is stable and has random chaos.(2)Synchronization of mobile communication time series.The synchronization of large traffic events between sectors may lead to system congestion.Based on event synchronization and network modeling,the synchronization of large traffic events in mobile communication is studied.The event synchronization method is used to extract the large traffic event sequence from the sector data traffic throughput time series.The event synchronization network is constructed by network modeling method,and the node congestion coefficient is constructed according to the local topological characteristic parameters of the network.The congestion coefficient considers not only the frequency of synchronization events,but also the probability of synchronization events among the topological neighbors of the sector.The congestion coefficient of nodes quantifies the risk of congestion locally.The research finds that only a few sectors will have synchronous large traffic events many times;Overall,the risk of congestion is low.(3)The coupling of mobile communication time series.Based on the data throughput and traffic time series of mobile communication,the data voice coupling binary network is constructed according to the sliding correlation method and Spearman rank similarity to reveal the dynamic change of data traffic and voice traffic coupling.The cumulative degree distribution of the network obeys the linear distribution.This indicates that the coupling between data traffic and voice traffic is uniform.The core-periphery model is used to analyze the spatiotemporal clustering characteristics of the coupling between throughput and traffic.The coupling proportion is defined to visualize the core area and the periphery area.The results of spatiotemporal clustering show that the frequency of data throughput and traffic in the core area is higher than that in the periphery area.