MOOC中基于二分图推荐的同伴互评系统优化

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；接着，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。

Peer review system optimization based on bipartite graph recommendation in MOOC

On the question of homework correcting not in time in MOOC, it 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. Builds the model of reviewer recommendation. Secondly, it designs 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.