User Preference Learning Based Proactive Content Caching in Mobile Edge Computing
Shuailong Cao, Qian Zhang, Yusong Lin · 2023
With the widespread deployment of 5G networks, wireless network traffic has witnessed a sharp increase. Leveraging Mobile Edge Computing (MEC) technology to cache content at the network edge can effectively reduce the load on network backhaul traffic and provide users with low-latency, high-bandwidth services. However, existing proactive caching solutions mostly rely on content popularity, often neglecting user preferences. Therefore, this paper introduces a user preference learning based proactive caching strategy (UPLPC). Specifically, it utilizes convolutional collaborative filtering to capture user preferences and enhances the network’s expressiveness through a channel attention mechanism. Subsequently, a fully connected neural network is utilized to predict user preferences for different content. Experimental results indicate that UPLPC can enhance the accuracy of user preference prediction and increase cache hit ratio on edge servers.