The reconstruction on the game networks with binary-state and multi-state dynamics.

Area of Science:

• Complex systems analysis
• Network science
• Data inference

Background:

• Network reconstruction is crucial for understanding complex systems but suffers from low accuracy with small datasets.
• Threshold subjectivity in adjacency matrix inference poses a significant challenge.

Purpose of the Study:

• To propose novel methods for accurate network reconstruction using limited observation data.
• To address the limitations of existing techniques in handling small data and threshold subjectivity.

Main Methods:

• Development of Quadratic Compressive Sensing (QCS) and Integer Compressive Sensing (ICS) models.
• Introduction of a Combined Compressive Sensing (CCS) method integrating QCS and ICS.
• Application to binary-state and multi-state network dynamics.

Main Results:

• CCS demonstrates superior performance over standard compressive sensing and LASSO across various network structures and scales.
• CCS achieves higher accuracy in inferring larger nodes compared to other methods.
• The proposed methods effectively reveal hidden relationships within complex systems using minimal data.

Conclusions:

• The CCS method provides a robust approach for network reconstruction, particularly effective with limited data.
• This work enhances the ability to understand, predict, and control intricate systems.
• The findings contribute to advancing network science and complex systems analysis.