摘要
在“互联网+”在线学习环境下,探索符合学习习惯和偏好的个性化推荐路径能够降低学习者学习的盲目性、提升学习者在线学习体验。文章提出了一种基于学习者画像的个性化课程推荐方法,首先,利用爬虫技术获得Bilibili网站30多万名学习者的数据,然后对学习者学习数据进行定量分析,尤其是在个性化特征最明显的情感表达方面,采用了基于注意力机制的双向长短时记忆网络进行情感分析,从而构建了包含学习者基本信息、行为和弹幕文本三个维度的学习者画像特征模型。在此基础上,利用深度神经网络建立了教学资源与学习者画像之间的关系模型,用该模型预测学习者新的学习需求。实验结果表明,当学习者登录后,输入要选择的课程,模型能够根据学习者画像推荐相似学习者学习过的课程,提供个性化课程推荐服务,且推荐评价指标也表明该模型能够提高推荐性能。
In the"Internet+"online learning environment,exploring personalized recommendation paths in line with learning habits and preferences can reduce the blindness of learners and improve their online learning experience.This paper proposes a personalized course recommendation method based on learners'portraits.Firstly,the crawler technology is used to obtain the data of more than 300000 learners on Bilibili website.Then,based on the quantitative analysis of the learners'learning data,a learner's portrait feature model is constructed,which includes learner's basic information,behavior and bullet-screen text.Especially,when it comes to the expression of emotion,as the most obvious individualized feature,a bidirectional short and long-term memory network based on attention mechanism is used for emotion analysis.On this basis,the relationship model between teaching resources and learners'portraits is established by using deep neural network,and the model is used to predict learners'new learning needs.The experimental results show that when the learner logs in and enters the courses to select,the model can recommend courses learned by similar learners according to their portraits,and provides personalized course recommendation service.The recommendation evaluation index also shows that the model can improve the recommendation performance.
作者
王莉莉
郭威彤
杨鸿武
WANG Lili;GUO Weitong;YANG Hongwu(Advanced Innovation Center for Future Education,Beijing Normal University,Beijing 100875;School of Educational Technology,Northwest Normal University,Lanzhou Gansu 730070)
出处
《电化教育研究》
CSSCI
北大核心
2021年第12期55-62,共8页
E-education Research
基金
2020年国家自然科学基金地区项目“藏族地区儿童国家通用语言口语智能学习的研究”(项目编号:62067008)
甘肃省教育科学“十三五”规划2020年度重点课题“面向东乡族的国家通用语学习关键技术的研究”(课题编号:GHBZ190)。
关键词
学习者画像
深度神经网络
个性化推荐
非正式学习平台
教育大数据挖掘
Learner Portrait
Deep Neural Network
Personalized Recommendation
Informal Learning Platform
Big Data Mining for Education
