Multi-source detection based on neighborhood entropy in social networks

Area of Science:

• Computer Science
• Network Science
• Information Science

Background:

• Social networking platforms facilitate the rapid dissemination of false information.
• Existing source location methods are often ineffective due to limited use of neighborhood information and failure to account for multiple sources.

Purpose of the Study:

• To develop an effective method for locating multiple sources of false information in complex networks.
• To improve the comprehensiveness and accuracy of source localization compared to existing approaches.

Main Methods:

• Proposed a new multiple source location method utilizing neighborhood entropy.
• Defined infection adjacency entropy and infection intensity entropy to evaluate node infection possibility.
• Developed a source location algorithm for infected clusters to identify multiple sources using node cohesion.

Main Results:

• The proposed methods demonstrate superior performance over existing techniques in source localization experiments.
• Experimental results were validated across various network topologies, confirming the effectiveness of the approach.

Conclusions:

• The novel neighborhood entropy-based method significantly enhances the accuracy and comprehensiveness of false information source localization.
• This approach effectively addresses the challenge of identifying multiple sources within infected clusters on social networks.