KNN based Novel Recommendation System for Mobile Edge Caching

Edge caching is a key feature in enabling a robust wireless system to reduce the backhaul network traffic. In this article, a novel cache management based on recommendation systems is developed based on user preference. By considering user’s movie consumption, we use the user’s rating of the movie to maintain the cache, so as to reduce backhaul network latency. A simple unsupervised KNN based algorithm has been proposed to predict the movie ratings in a edge caching system. A strategy is proposed to maintain data in a cache in such a way that maximizes the number of times data requested by the user is present in the cache. The proposed algorithm is compared with some baseline algorithms. The simulation results show our proposed algorithm has a gain of 30.7% average cache hit compared to the baseline algorithms and an improvement of 2.3% in delay.