A node influence ranking algorithm combining k-shell iteration and node degree.

Area of Science:

• Network Science
• Computational Social Science
• Data Mining

Background:

• Identifying key nodes is crucial for understanding information/disease spread in complex networks.
• Traditional K-shell decomposition lacks discrimination due to reliance on global position only.
• Existing methods fail to simultaneously use iteration factor and degree for node distinction.

Purpose of the Study:

• To propose a novel node influence ranking algorithm.
• To enhance K-shell decomposition by integrating global and local network topology information.
• To improve the accuracy and discrimination of key node identification in complex networks.

Main Methods:

• Developed a new algorithm integrating K-shell iteration, node degree, and neighbor information.
• Evaluated the algorithm through simulation experiments on eight diverse networks.
• Compared performance against established methods like dc, bc, cc, k-shell, Ks+, KSIF, LGI, and DCK.

Main Results:

• The proposed algorithm demonstrated superior accuracy in ranking key nodes compared to existing methods.
• Achieved an average accuracy improvement of 5.15% over the second-best algorithm.
• The KTD algorithm showed higher accuracy in identifying top key nodes and strong discriminative power.

Conclusions:

• The integrated approach offers enhanced accuracy and discriminative power for key node identification.
• The algorithm maintains good time performance, making it suitable for large-scale complex networks.
• This method provides a more robust solution for network analysis applications.