基于标签与评分个性化课程推荐算法

Personalized curriculum recommendation algorithm based on tags and ratings

首先根据课程的标签对课程进行聚类,找到相似度高的课程;其次根据学生对课程的已有评价和课程的聚类结果对未选课程进行评分预测,构建无缺失的课程评价矩阵,在此基础上再次对课程进行相似度计算,找到相似度较高的K项向目标用户进行推荐。通过实验验证,本算法与基于标签协同过滤算法以及基于评分推荐算法相比,具有更准确的推荐效果。

Cluster analysis is carried out according to the tags of the curriculum to find the high similarity of courses.Then rating prediction for a course is completed based on the previous evaluation and cluster analysis,and the intact curriculum evaluation matrix is established.Again,similarity of the courses is calculated to find Kitems for recommendation to the target users.Experiments indicate that the algorithm is more accurate for recommendation than those based on tags or rating.