基于协同标注的个性化电子学习推荐系统

Personalized electronic learning recommendation system based on collaborative tagging technology

分析各种基于标注的推荐技术的适用性,通过张量因子分解技术对最佳模型排序方法进行改进;比较基于标注的协同过滤(CF)、基于图的方法(自适应Page-Rank和Folk-Rank算法)以及基于张量的方法(HOSVD和RTF)的预测性能,优选出最佳的推荐方案。通过在一个开发的编程辅导系统中的应用,验证了所提推荐系统的有效性。

The applicability of various tag-based recommendation techniques was analyzed,the best model ranking method of tensor factor decomposition technique to obtain the best recommendation effect was improved,and the predictive performances of annotation-based collaborative filtering(CF),graph-based methods(adaptive Page-Rank and Folk-Rank algorithms),and tensor-based methods(HOSVD and RTF)were compared,and the best recommended scheme was selected.Through the application in a programming tutoring system developed,the effectiveness of the proposed method was verified.