基于LDA模型的高校教育云平台设计与研究

Design and Research of University Education Cloud Platform Based on LDA User Interest Model

随着互联网技术的飞速发展,参与在线学习的群体日益增多,在线学习、互动和交流的活动越来越频繁,在线教育作为一种新型的教育模式逐步形成。一方面,它给学生带来了极大的方便,并且提供了一种新的学习方式;另一方面,也对学习资源的快速增长带来的“信息过载”现象提出了解决方案。文中在云计算环境中采用基于分层的体系化设计思路,深入开展了高校教育云的需求分析和架构设计。结合Dubbo微服务框架和分布式数据库技术,以服务化、组件化为系统核心,基于LDA用户兴趣模型分析用户对教育主题的偏好程度,确定主题与在线教育课程的相似度系数,从而实现对用户的教育课程精准推送。系统上线运行结果表明,使用基于LDA模型的课程推荐较传统的推荐方法更为精准,主题集中度也更高,更加适合在线的教育课程推荐。

With the rapid development of Internet technology,more and more people participate in online learning.Online learning,interaction and communication activities are more and more frequent.As a new education mode,online education has been gradually formed.On the one hand,it brings great convenience to students and provides a new way of learning.On the other hand,it also provides solutions for the phenomenon of"information overload"brought by the rapid growth of learning resources.In the cloud computing environment,the demand analysis and architecture design of college education cloud are carried out in depth with the idea of hierarchical systematization design.Combined with Dubbo micro-service framework and distributed database technology,with sertization and componentization as the core of the system,based on LDA user interest model,the preference degree of users for educational topics is analyzed,and the similarity coefficient between topics and online education courses is determined,so as to achieve accurate push of educational courses to users.The results of online operation show that the LDA based course recommendation method is more accurate than the traditional recommendation method,and the topic concentration is higher,so it is more suitable for online education curriculum recommendation.