Modelling and monitoring social network change based on exponential random graph models

Area of Science:

• Social Network Analysis
• Statistical Modeling
• Real-time Data Analysis

Background:

• Social networks exhibit dynamic structures that can change over time.
• Detecting structural anomalies in real-time is crucial for understanding network evolution and potential disruptions.
• Existing methods may lack efficiency for continuous, real-time monitoring of complex network changes.

Purpose of the Study:

• To develop a real-time method for detecting anomalous changes in social network structure.
• To provide early warning systems for significant structural shifts in evolving networks.
• To apply and validate the proposed methods on real-world social network data.

Main Methods:

• Utilizing the exponential random graph model (ERGM) for social network representation.
• Developing an online monitoring technique based on a split likelihood-ratio test for ERGMs.
• Employing pseudo-maximum likelihood estimation and a bisection algorithm for control limits.

Main Results:

• The proposed method effectively detects anomalous structural changes in simulated and real social network data.
• Performance evaluation demonstrates the procedure's sensitivity to different change points and shift sizes.
• The approach offers a computationally efficient alternative to Markov Chain Monte Carlo methods for online monitoring.

Conclusions:

• The developed ERGM-based online monitoring technique provides a robust framework for real-time anomaly detection in social networks.
• This approach facilitates early warnings of structural shifts, applicable to various dynamic network analysis scenarios.
• The study validates the method's efficacy through simulations and a real-world application on a social proximity network.