基于Hadoop-Mahout的分布式课程推荐算法  

DISTRIBUTED COURSE RECOMMENDATION ALGORITHM BASED ON HADOOP-MAHOUT

作　　者：徐文健 刘青昆[1] 郑晓薇[1] 李永波 Xu Wenjian;Liu Qingkun;Zheng Xiaowei;Li Yongbo(College of Computer and Information Technology,Liaoning Normal University,Dalian 116081,Liaoning,China;Wafangdian No.8 Senior High School,Dalian 116300,Liaoning,China)

机构地区：[1]辽宁师范大学计算机与信息技术学院,辽宁大连116081 [2]瓦房店市第八高级中学,辽宁大连116300

出　　处：《计算机应用与软件》2018年第3期236-240,共5页Computer Applications and Software

基　　金：国家自然科学基金项目(61373127)

摘　　要：针对MOOC(Massive Open Online Course)平台上同类及相似课程繁杂,在线学习者不易找到适合自己的课程,而导致学习效率降低,学习效果较差等问题,提出一种IRS课程评价方法,对在线课程进行相关的评价。结合用户偏好及IRS方法改进了机器学习框架Apache-Mahout的协同过滤推荐算法,对在线学习者进行个性化课程推荐。面对MOOC平台上大量的课程信息及学习者信息,基于Hadoop分布式云计算平台,设计了在线课程推荐并行算法。实验结果表明,提出的IRS推荐算法有效且适用于分布式云计算环境,同时验证了该算法在分布式环境下并行计算的高效性。In view of the complicated and similar courses on the MOOC platform,online learners are not easy to find their own courses,which leads to the problems of low learning efficiency and poor learning effect.The article proposed a method of IRS course evaluation.Machine learning framework Apache-Mahout collaborative filtering recommendation algorithm was improved with user preference and IRS methods to recommend online learners personalized courses.Faced with a large number of course information and learner information on the MOOC platform,an online course recommendation parallel algorithm was designed based on the Hadoop distributed cloud computing platform.The experimental results showed that the proposed IRS recommendation algorithm proposed in this paper was effective and suitable for distributed cloud computing environment.Meanwhile,the efficiency of parallel computing in distributed environment was verified.

关 键 词：MOOC IRS课程评价 推荐算法 Mahout HADOOP 

分 类 号：TP302[自动化与计算机技术—计算机系统结构]