A network method to analyze compound extreme events: Risk enhancement relationship and trigger causal relationship in high voice traffic and high data throughput events

Abstract

Based on ideas from event coincidence analysis (ECA), we propose a network analysis method to study compound extreme events at different geographical locations. Integrating network modeling into statistical correlation research allows us to analyze potential risk enhancement relationship and trigger causal relationship between these events. In this approach, we consider different geographical locations as nodes and construct a directed edge from node i to node j when event A at location i occurs synchronously before event B at location j. Precursor coincidence analysis quantifies the risk enhancement relationship between two types of extreme events, while trigger coincidence analysis quantifies the trigger causal relationship between two types of extreme events. A directed weighted network can be constructed based on statistical correlations between these events at different geographical locations. Further analysis of network topology characteristics extends traditional ECA in method and application. Herein, we construct the precursor functional network and the trigger functional network of high voice traffic and high data throughput to analyze potential risk enhancement and trigger causal relationships between these events at different base stations within a communication system.