With the maturity of high bandwidth and low latency of 5G network, social media data presents explosive growth and highly dynamic trend, which brings severe challenges to the traditional data processing architecture. Focusing on the real-time perception and multi-source fusion of social media data streams in 5G environment, this paper proposes a three-tier system architecture based on edge-cloud collaboration, which combines the adaptive dynamic sampling strategy with the fusion algorithm based on information entropy and reinforcement learning (RL) to realize efficient processing and intelligent fusion of multi-modal data. The experimental results show that the proposed Edge-RL-Fusion algorithm is significantly superior to the traditional methods in key event information retention rate, fusion accuracy and system response delay, and has good real-time and robustness. This study provides a feasible technical path for intelligent perception and fusion of social media data in the 5G environment, and has important theoretical value and practical significance for smart cities, public opinion monitoring and other application scenarios.
Riming Zhou (Sun,) studied this question.