协同过滤推荐算法在视频缓存策略中的应用

Application of collaborative filtering recommendation algorithm in video caching strategy

缓存替换策略是内容分发网络(Content Delivery Network,CDN)研究中的重要内容。常用的缓存方案是根据内容本身的特征进行缓存,比如视频的流行度、评分质量等。文章在缓存策略的设计中考虑到了用户特征,即根据用户的喜好,选择需要缓存的内容。使用基于矩阵分解的推荐算法对用户需求进行分析,筛选出用户可能感兴趣的视频,并利用基于加权评分预测值的贪婪缓存算法选择合适的内容进行缓存。仿真实验的结果表明,该算法可以将缓存命中率提高5-10%。

The cache replacement strategy is an important content in the research of content delivery network(CDN).The caching scheme commonly used is to cache according to the characteristics of the content itself,such as the popularity of the video,the quality of the rating,and so on.This paper takes into account the characteristics of users in the design of the caching strategy,that is,select the content that needs to be cached according to the user's preferences.A recommendation algorithm based on matrix factorization is used to analyze users'needs,and screen out the videos that users may be interested in,and use a greedy caching algorithm based on weighted score predictions to select appropriate video content for caching.The results of simulation experiments show that the algorithm can increase the cache hit rate by 5-10%.