Topic-aware influence maximization based on deep reinforcement learning and group relative optimization

Key Points

• Influence maximization improves by 25% through the proposed algorithm enhancements with deep reinforcement learning.
• Key metrics show significant gains in user engagement patterns and network reach across multiple scenarios.
• Analysis focused on algorithmic performance in various social network structures, evaluating interaction patterns effectively.
• The findings highlight the importance of tailored strategies in maximizing influence, with future applications in marketing and social dynamics.