Reconstructing Sparse Multiplex Networks with Application to Covert Networks.

Area of Science:

• Network science
• Complex systems analysis
• Data mining and machine learning

Background:

• Understanding complex systems relies heavily on network structure.
• Complete network structures are often unavailable in real-world scenarios.
• Inferring network structures is crucial for system analysis and prediction.

Purpose of the Study:

• To develop and validate a novel framework for reconstructing complete multiplex network structures.
• To compare the performance of the proposed framework against existing methods.
• To explore applications in monitoring covert networks and resource allocation.

Main Methods:

• Integration of the configuration model for random network generation within an Expectation-Maximization-Aggregation (EMA) framework.
• Validation using real-world multiplex networks (covert and overt).
• Comparative analysis against the Expectation-Maximization (EM) framework and a standard random model.

Main Results:

• The EMA framework generally achieved the highest predictive accuracy in reconstructing multiplex network structures.
• Performance improvement of EMA over EM decreased as the number of network layers increased.
• The method proved effective across various real-world network types.

Conclusions:

• The EMA framework offers a robust approach for inferring complete multiplex network structures.
• Inferred network structures can guide decisions in covert network monitoring and resource allocation.
• This method provides valuable insights for law enforcement in interdiction strategies.