Identifying Influential Nodes Based on Evidence Theory in Complex Network.

Area of Science:

• Complex network science
• Data analysis
• Network theory

Background:

• Identifying influential nodes is crucial in complex network science.
• Classical methods struggle with complex, high-dimensional real-world networks.
• Existing approaches often fail to adequately handle uncertainty and conflicting information.

Purpose of the Study:

• To propose a novel method for influential node identification using Dempster-Shafer (DS) evidence theory.
• To enhance the efficiency and reliability of influential node detection in complex networks.
• To demonstrate the method's effectiveness in network disintegration and financial time series analysis.

Main Methods:

• The proposed method leverages Dempster-Shafer (DS) evidence theory.
• DS theory quantifies uncertainty using basic belief assignment functions.
• Dempster's rule of combination is employed to process conflicting evidence and integrate multidimensional information.

Main Results:

• The DS method significantly improves influential node identification compared to classical algorithms.
• Attacking nodes identified by the DS method leads to greater network disintegration.
• Application to GBP futures time series reveals DS method identifies key price turning points.

Conclusions:

• The Dempster-Shafer (DS) evidence theory offers a robust framework for influential node identification in complex networks.
• The proposed method enhances network analysis reliability and provides insights into financial market dynamics.
• This approach effectively handles uncertainty and multidimensional data, outperforming existing techniques.