Machine Learning based Small Cell Cache Strategy for Ultra Dense Networks

Caching contents at base stations (BSs) has emerged as an effective way to offload backhaul traffic and improve quality of experience (QoE). Considering the limited cache size, how to maximize cache efficiency has become an urgent issue to be addressed. In this paper, we consider caching selected contents at small cell base stations (SBSs) in ultra dense network (UDN). The cache efficiency problem is formulated as a system backhaul load minimization problem, which is hard to be solved for the highly random content demands. Therefore, machine learning based cache strategies are proposed to tackle this difficult problem from the perspective of exploiting the potential of mobile traffic data. First, K-means clustering algorithm is used to fully uncover hidden spatio-temporal patterns of content requests at SBSs, and achieve personalized inter-cluster cache and predictive intra-cluster cache. Second, k-Nearest Neighbour (k-NN) classification algorithm is introduced to categorize the constantly emerging new contents and cache them in the corresponding cluster periodically with high accuracy and low complexity. Simulation results demonstrate great superiority of the proposed cache strategies over the existing approach.