个性化推荐技术的发展有利于解决互联网海量数据信息过载问题。本文在了解分布式处理架构的特点的基 础上，提出了基于Hadoop 构建标签推荐系统方案。采用基于MapReduce 模型实现的算法具有较高的伸缩性和性能，能高效 地进行离线数据分析。

Personalized recommendation technology development is conducive to resolve the Internet data information overload problem. Based on the understanding of the characteristics of distributed processing architecture, this paper proposes the tag recommendation system based on Hadoop. The algorithm based on MapReduce model has high scalability and performance, and can analyze offline data efficiently.