基于多维用户模型的数字教育服务推荐技术研究

摘要

目前的数字教育推荐系统中用户建模数据来源单一,用户意图分析不深入,缺乏数字教育的针对性考虑,准确度和满意度均有所局限.提出一种基于多维用户模型的数字教育服务推荐方法,充分考虑学生知识掌握水平、学习目标、学生所在群体等因素,从客观意图、主观意图和群体性意图三个方向进行分析推理.结合项目需求,开发基于上述推荐方法的数字教育应用原型系统.实验结果表明,该方法相对于仅考虑单一用户因素的系统在准确率、召回率及排序优先度等指标上均有明显提升.%Currently for e-learning recommendation systems,the user modelling data source is single,the user intention analysis is not deep,the lack of digital education~ pertinence consideration,the accuracy and the satisfaction degree are limited.This paper proposes a recommendation method of e-learning services based on multi-dimensional user model,taking into account the factors such as students' mastery level of knowledge,learning goals and students'groups,and reasoning from objective intention,subjective intention and group intention.Based on the requirement of the project,a prototype system of e-learning application based on the recommendation method is developed.The experimental results show that the proposed method can improve the accuracy,recall rate and sort priority of the system with only a single user factor.