MOOC中基于二分图推荐的同伴互评系统优化  

Peer review system optimization based on bipartite graph recommendation in MOOC

作　　者：何升[1] 邓伟林[2] 肖体斌 

机构地区：[1]广东轻工职业技术学院计算机系,广州510300 [2]中山大学信息科学与技术学院,广州510006 [3]广东工业大学机电工程学院,广州510006

出　　处：《计算机应用研究》2016年第5期1399-1402,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(61033010);国家科技重大专项基金资助项目(2008ZX10005-013);广东省教育厅"十二五"规划项目信息技术专项基金资助项目(13JXN051)

摘　　要：针对MOOC课程平台中作业反馈不及时的问题,就同伴互评系统的准确性与可靠性优化进行了研究。为了达到为作业推荐合适评阅人的目的,在综合考虑作业评阅人的评阅意愿、评阅能力和评阅双方作业相似度等多种因素的基础上,建立了作业评阅人推荐模型;引入二分图匹配理论来求解作业评阅人的任务均衡问题,设计了与之相适应的最优均衡匹配算法。实验表明,算法在评阅人的工作量均衡和作业评阅人的准确推荐等方面均取得了较好的效果,优化提高了同伴互评系统的准确性和可靠性。采用经互评算法优化的同伴互评系统,可以改善MOOC平台的整体满意度。On the question of homework correcting not in time in MOOC,this paper researched how to improve the accuracy and reliability of peer review system. In order to recommend right reviewer for every homework,integrated into account the willingness and ability of reviewer as well as homework similarity and other factors. It built the model of reviewer recommendation. Secondly,it designed an optimal balanced matching algorithm with modified bipartite graph matching method to meet the need of reviewer work balancing and recommending right reviewer for every homework. Experimental results show that,results about reviewer work balancing and reviewer recommendation accurately are satisfactory. It can improve the overall satisfaction of MOOC platform through apply peer algorithm to optimize peer reviewer system.

关 键 词：大规模开放在线课程 二分图 同伴互评 推荐 均衡分配 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] TP182[自动化与计算机技术—计算机科学与技术]