社交网络大数据下贪婪式实时网站推荐算法  

User-based greedy real-time websites recommendation algorithm for big data in social network

作　　者：娄建楼[1] 邹伟[1] 王玲[1] 曲朝阳[1] 史春雷[1] 

机构地区：[1]东北电力大学信息工程学院,吉林吉林132012

出　　处：《计算机应用研究》2015年第5期1361-1364,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(51277023);吉林省科技发展计划项目(20150204084GX);吉林市科技计划项目(201414011)

摘　　要：社交网络每天都会产生结构化、半结构和非结构化的大数据,数据的增长速度超过了对硬件需求的摩尔定律。在社交网络中还存在各种恶意评价、刷分和刷网站关注度等不良现象,对大数据的分析处理带来了巨大挑战。为了提高数据的处理效率和网站推荐的准确性,提出了一种在Hadoop云平台下基于用户的贪婪式实时近似网站推荐的RT-G算法。算法通过迭代寻优算法找到最合适的用户数量作为网站推荐评价的用户标准,应用频度近似算法完成对网站的推荐,通过实验证明了方法的效率和有效性。The social network will produce big data of structured,semi-structured and unstructured every day,and the growth of the data exceeds the demand for hardware of Moore＇s law. Especially there are some bad phenomenon such as various malicious evaluation,or getting points and attention by unfair means,which is a challenge for big data of processing and analyzing.In order to improve the accuracy of the data processing efficiency and website recommendation,this paper proposed a new method that named RT-G algorithm to solve this problem in a Hadoop cloud platform. It found the most suitable number of users as the evaluation of user standard of the websites recommendation applying iterative optimization algorithm as the first step,and then the frequency approximation algorithm completed the recommendation of websites by the experimental analysis of real data sets,at last the experiment proves the efficiency and effectiveness of the method.

关 键 词：社交网络大数据 HADOOP 网站推荐 实时 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] TP181[自动化与计算机技术—计算机科学与技术]