【作者】 王新；

【导师】 王素琴； 郭宗一；

【作者基本信息】 华北电力大学(北京) ， 工程硕士（专业学位）， 2018， 硕士

【摘要】 随着互联网的迅速发展,在线学习逐渐成为一种被大众广泛认可并采用的学习方式。但是,大量学习资源的出现,使得学习者无法快速发现自己感兴趣的学习资源,这是迫切需要解决的问题。个性化推荐技术是根据学习者的学习行为,主动为学习者推荐恰当的学习资源的推荐技术,它在解决学习资源超载问题和满足不同学习者的需求方面已经得到研究者的高度重视。协同过滤算法是以学习者对学习资源的评分数据为依据,向不同的学习者推荐相似的学习资源,分为基于用户的协同过滤算法和基于项目的协同过滤算法。通过实验发现在学习资源推荐方面,基于项目的协同过滤算法准确率更高。同时,考虑到在学习过程中涉及的学习资源具有一定的时序性的特点,将频繁模式挖掘技术应用于学习资源推荐中。GSP算法能够快速地挖掘出学习者的最大频繁学习资源序列,向不同的学习者推荐接下来最可能需要的学习资源。单一的推荐算法无法得到满意的推荐结果,本文采用将协同过滤算法和GSP算法的推荐结果相结合的混合推荐算法。通过数据集验证,这种混合推荐算法,F指标达到34%左右,与单独使用一种推荐算法相比,有了明显的提高。最后实现了学习资源个性化推荐系统,将基于项目的协同过滤算法和GSP算法结合的混合推荐算法应用到该系统中,实现了学习资源的精准推荐,效果良好。更多还原

【Abstract】 With the rapid development of the Internet,the online learning has gradually become a way of learning,which is widely recognized and adopted by the public.With the rapid increase of learning resources in the Internet,it is an urgent problem to help the learners find the learning resources that they are interested in quickly.Personalized recommendation technology is a recommendation technology,which recommends appropriate learning resources for learners based on learners’ learning behavior.It has been valued by researchers in solving the problem of overloading learning resources and meeting the needs of different learners.Collaborative filtering algorithm is based on learners’ score data of learning resources,recommends similar learning resources to different learners,and divides them into user-based collaborative filtering algorithm and item-based collaborative filtering algorithm.Through experiments,it is found that item-based collaborative filtering algorithm is more accurate in the learning resource recommendation.At the same time,considering that learning resources involved in the learning process have certain timing characteristics,frequent pattern mining technology is applied to the learning resource recommendation.The GSP algorithm can quickly excavate the learners’ maximum frequent learning resources sequence and recommend learning resources that are most likely to be needed for different learners.Single recommendation algorithm can not get the satisfactory recommendation results.In this paper,the Hybrid recommendation algorithm based on collaborative filtering algorithm and GSP algorithm recommendation results is adopted.Through data set verification,this hybrid recommendation algorithm has a F index of about 34%,which is significantly improved compared with a single recommendation algorithm.Finally,the personalized recommendation system of learning resources is implemented.The hybrid algorithm based on item-based collaborative filtering algorithm and GSP algorithm is applied to the system,which achieves accurate recommendation of learning resources and good results.更多还原