基于TFIDF+LDA和Mini Batch K⁃means算法的在线课程推荐方法研究  

Research on online course recommendation method based on TFIDF+LDA and Mini Batch K⁃means algorithm

作　　者：严武军[1] 王丽蓉 Yan Wujun;Wang Lirong(College of Computer Science and Technology,Taiyuan Normal University,Jinzhong 030619,China)

机构地区：[1]太原师范学院计算机科学与技术学院,晋中030619

出　　处：《现代计算机》2023年第23期15-20,共6页Modern Computer

摘　　要：在线教育资源急剧增长让学习者难以抉择,研究在线课程分类推荐,能帮助学习者快速获取所需资源。首先将潜在狄利克雷分配算法融入词频-逆向文件频率算法对数据进行预处理,生成词向量矩阵;之后采用Mini Batch K-means算法训练聚类模型,并采用T分布随机邻域嵌入降维算法对训练结果进行可视化分析。实验采用从Pluralsight在线课程API获取8016条数据进行实验,实验结果表明融入潜在狄利克雷分配算法的词频-逆向文件频率算法效果更好。The rapid growth of online education resources makes it difficult for learners to choose,and the study of online course categorization recommendation can help learners quickly access the required resources.Firstly,the potential Dirichlet allo-cation algorithm is integrated into the word frequency-inverse file frequency algorithm to preprocess the data and generate the word vector matrix;after that,the Mini Batch K-means algorithm is used to train the clustering model and the T-distributed stochastic neighborhood embedding dimensionality reduction algorithm is used to visualize and analyze the training results.The experiments are conducted by obtaining 8016 data from Pluralsight online course API,and the experimental results show that the word frequency-inverse file frequency algorithm incorporating the potential Dirichlet allocation algorithm is more effective.

关 键 词：词频逆向文件频率 潜在狄利克雷分配 Mini Batch K-means 在线课程推荐 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术] G434[自动化与计算机技术—计算机科学与技术]