Proactive Caching Strategy With Content-Aware Weighted Feature Matrix Learning in Small Cell Network

Abstract

Content caching in wireless edge is considered to be a feasible approach to improve the user’s quality of experience by caching popular multimedia content from data center. In this letter, we focus on movie on-demand streaming from Internet-based servers and propose a novel proactive caching strategy over time in small cell network. First, it constructs a weighted feature matrix (WFM) and estimates content popularity in terms of spatiotemporal variation using content-aware WFM (CWFM). According to estimated content popularity, the CWFM-based proactive caching algorithm makes a caching decision by cooperatively caching the contents with a approach of linear programming and variable recovery to minimize the average delivery latency (ADL) in each time slot. The simulation results show that our algorithm improves the performance in terms of ADL based on MovieLens data set.