基于Hadoop平台的图书推荐方法研究  

Research on book recommendation technology based on Hadoop platform

作　　者：王晋月[1] WANG Jin⁃yue(Library,Shangluo Institute,Shangluo 726000,China)

机构地区：[1]商洛学院图书馆

出　　处：《电子设计工程》2019年第24期20-23,共4页Electronic Design Engineering

基　　金：商洛学院研究项目(18SKY019)

摘　　要：考虑到大数据环境下传统的单机推荐算法无法高速有效地处理大规模的数据,为了提高图书推荐算法性能和解决图书推荐系统可扩展性的问题,通过Hadoop平台下的MapReduce编程模型,提出一种分布式加权型混合推荐算法。在分布式环境下采用基于矩阵分解的协同过滤推荐算法,改善数据稀疏性,然后将聚类模型结合矩阵分解算法,解决读者数据冷启动问题。在Hadoop上实现该算法不仅提升了算法的运行速率,而且能够有效解决算法的可扩展性问题。In order to improve the performance of book recommendation algorithm and solve the problem of scalability of book recommendation system.other than that the traditional single-machine recommenda⁃tion algorithm can not deal with large-scale data efficiently in the environment of big data.Based on the MapReduce programming model under the Hadoop platform,a distributed weighted hybrid recommenda⁃tion algorithm is proposed.In the distributed environment,the collaborative filtering recommendation algorithm based on matrix decomposition is used to improve the sparsity of data,and then the cluster model is combined with matrix decomposition algorithm to solve the problem of cold start of reader data.Implementing the algorithm on Hadoop not only improves the running speed of the algorithm,but also effectively solves the problem of the scalability of the algorithm.

关 键 词：Hadoop系统 混合推荐 并行计算 图书推荐 

分 类 号：TN49[电子电信—微电子学与固体电子学]