摘要
随着互联网的高速发展,海量新闻的个性化推荐成为一个重要课题,针对海量新闻的个性化推荐算法进行研究,以MapReduce的并行方法设计了基于Hadoop云平台下的协同过滤算法,利用MapReduce的并行方法,将传统的协同过滤算法并行化,并详细说明了并行化步骤和实现细节;最后用实验结果验证了改进的并行化的协同过滤算法在运行速度和执行效率方面有明显的提高,更适合处理大数据。
With the rapid speed development of Internet, personalized recommendation for massive news has become an important sub ject. The research focus on the personalized recommendation algorithm for massive news. By means of MapReduce parallel method a collaborative filtering algorithm based on Hadoop distributed platform is proposed. Using MapReduce parallel method, The paper illuminates that the traditional collaborative filtering algorithm is parallelized and how to realize the parallel steps and implementation details of the algorithm. At last experimental results show that the improved parallel collaborative filtering algorithm has improved obviously in the running speed and execution efficiency, is also much suitable for processing big data.
出处
《计算机测量与控制》
2015年第6期2082-2085,共4页
Computer Measurement &Control
基金
国家自然科学基金(31101085)
关键词
推荐系统
大数据
并行计算
协同过滤
recommendation system
big data
parallellzation
collaborative filtering
