Susceptible user search for defending opinion manipulation

Area of Science:

• Cybersecurity
• Social Network Analysis
• Artificial Intelligence

Background:

• Cyberspace enables convenient online communication, but also facilitates malicious manipulation of user opinions.
• Targeting influential and susceptible users is a common tactic for controlling public opinion, posing security threats.
• Existing defense strategies often lack efficiency or guaranteed performance in real-world scenarios.

Purpose of the Study:

• To develop an intelligent and efficient searching strategy for identifying susceptible and key users for defense against malicious manipulation.
• To address the limitations of current studies that offer solutions only for ideal scenarios or provide inefficient methods.

Main Methods:

• A greedy algorithm is proposed, specifically considering user susceptibilities.
• Unsupervised learning and community properties are utilized to design an accelerated algorithm.
• Approximation guarantees for both greedy and community-based algorithms are systematically analyzed.

Main Results:

• The proposed greedy and community-based algorithms demonstrate effectiveness in discovering susceptible users.
• Extensive experiments on real-world datasets show significant outperformance compared to state-of-the-art algorithms.
• The algorithms provide guaranteed performance in practical circumstances.

Conclusions:

• The combination of unsupervised learning and heuristic search offers a robust approach to defending against malicious online manipulations.
• The developed algorithms are efficient and effective in identifying targeted users for public security.
• This research contributes to enhancing cybersecurity by providing practical solutions for user manipulation defense.