摘要
为了解决MOOC平台课程推荐系统数据稀疏性的问题,该文提出一种基于稀疏偏好的矩阵分解和深度学习图像特征提取的混合推荐算法,用于提升推荐系统的质量。对MOOC推荐系统业务流程进行分析,并对基于深度学习的推荐系统架构和功能进行描述。利用深度学习技术,把课程封面视觉特征提取出来作为稀疏偏好矩阵的附加信息,提出了基于矩阵分解和深度学习融合的推荐方法。实验结果表明:该推荐算法具有较好的准确率,有效地缓解了数据稀疏问题。
In order to solve the problem of data sparsity in MOOC platform course recommendation system,a hybrid recommendation algorithm based on sparse preference matrix decomposition and deep learning image feature extraction is proposed to improve the quality of recommendation system.The business process of MOOC recommendation system is analyzed,and the architecture and function description of recommendation system based on deep learning are presented.By using the deep learning technology to extract the visual features of course cover as additional information of sparse preference matrix,a recommendation method based on matrix decomposition and deep learning is proposed.The experimental results show that the proposed algorithm has good accuracy and can effectively alleviate the problem of data sparsity.
作者
王艳
丁雪梅
孙薇
WANG Yan;DING Xuemei;SUN Wei(College of Pharmacy,Jilin University,Changchun 130012,China;College of Animal Science,Jilin University,Changchun 130012,China)
出处
《实验技术与管理》
CAS
北大核心
2020年第8期54-57,共4页
Experimental Technology and Management
基金
长春市科技发展计划资助项目(18YJ012)。
