Online Reconstruction of Complex Networks From Streaming Data.

Area of Science:

• Complex systems analysis
• Network science
• Data science

Background:

• Reconstructing nonlinear and complex dynamical systems is crucial across various scientific fields.
• Existing methods struggle with large-scale, real-time streaming data for network structure reconstruction.
• This limitation hinders real-time analysis and control of complex systems.

Purpose of the Study:

• To extend the network reconstruction problem (NRP) to online settings.
• To develop an effective method for online complex network reconstruction from streaming data.
• To enable real-time analysis and control of complex systems.

Main Methods:

• Extension of the network reconstruction problem (NRP) to online settings.
• Development of a follow-the-regularized-leader (FTRL)-Proximal style algorithm, termed Online-NR.
• Validation using synthetic evolutionary game network datasets and eight real-world networks.

Main Results:

• Online-NR successfully reconstructs network structure from large-scale real-time streaming data.
• The method demonstrates effectiveness in online network reconstruction tasks.
• Online-NR matches or surpasses the performance of nine state-of-the-art network reconstruction methods.

Conclusions:

• Online-NR addresses the limitations of current methods for online network reconstruction.
• The developed method enables effective real-time analysis and control of complex systems.
• Online-NR represents a significant advancement in handling large-scale, dynamic network data.