基于深度学习的课程资源推荐技术研究

Research of Course Resource Recommendation Technology Based on Deep Learning

随着互联网的发展,互联网+在不断的改变我们的传统产业,教育行业也是如此。在互联网+教育的线上教育模式的推动下,各种数字化工具的运用也加速了学校的数字化转型。学习者的学习模式不再局限于传统的线下学习,各种线上课程也被大家认可和喜爱,各种在线教育平台也逐渐得到推广。在线教育平台具有课程资源丰富、能实时开展在线学习的特点,得到了众多学习者的青睐。但随着课程资源的指数级增加,学习者难以从众多的课程中找到适合自己的课程,为了提升用户粘性,教育平台也引入了推荐算法。在传统的推荐算法中,比较常用的是协同过滤推荐算法,此算法在一定程度上能提升推荐的准确率。但协同过滤推荐算法也面临冷启动和数据稀疏的问题,为了解决此问题,本课题引入了基于深度学习的神经网络方法,将深度学习和协同过滤结合,进一步提升线上平台中课程资源推荐的准确率。

With the development of the Internet,Internet plus is constantly changing our tradi-tional industries,as is the cducation industry.Driven by the online education mode of Internet plus education,the use of various digital tools has also accelerated the digital transformation of schools.The learning mode of learners is no longer limited to traditional offline learning,and various online courses are also recognized and loved by everyone,and various online education platforms are gradually being promoted.The online education platform has the characteristics of rich course resources and real-time online learning,which has been favored by many learners.But with the exponential increase of course resources,learners find it difficult to find suitable courses from numerous courses.In order to improve user stickiness,education platforms have al-so introduced recommendation algorithms.In traditional recommendation algorithms,the com-monly used onc is the collaborative filtering rccommendation algorithm,which can improve the accuracy of recommendations to a certain extent.However,collaborative filtering recommenda-tion algorithms also face the problems of cold start and sparse data.In order to solve this prob-lem,this project introduces a neural network method based on deep learning,which combines deep learning and collaborative filtering to further improve the accuracy of course resource rec-ommendations on online platforms.