融合学习者社交网络的协同过滤学习资源推荐

摘要

传统的协同过滤推荐算法存在冷启动和数据稀疏的问题，使得新学习者因历史学习行为记录稀疏或缺失而无法获得较准确的个性化学习资源推荐。鉴于此，文章提出将学习者社交网络信息与传统协同过滤相融合的方法，计算新学习者与好友之间的信任度，借助新学习者好友对学习资源的评分数据，来预测新学习者对学习资源的评分值，以填补新学习者在学习者—学习资源评分矩阵中的缺失，实现对新学习者的个性化学习资源推荐。实证研究结果表明，该方法在一定程度上能够解决传统协同过滤方法的冷启动和数据稀疏问题，提高个性化学习资源推荐的准确率。%The traditional collaborative filtering recommendation algorithm (CF) has the problems of cold start and data sparseness, so it is difficult for new learners to obtain learning resources recommendation because of the sparsity or the lack of history learning behavior records. Thus, in this paper, a new CF method was proposed which combined learners’ social network information with traditional CF method. In order to fill the gap in the “learner - learning resource rating matrix” and achieve the recommendation of learning resources to new learners, it first calculated the trust degrees between new learners and their friends and then predicts new learners’ ratings on learning resources by virtue of that of their friends. The experimental results shown that this method can solve the cold start and data sparsity problems in some degree and improve the accuracy of personalized learning recommendation.