Article Preview
TopIntroduction
Currently, the course selection mode of most colleges is still releasing courses with students selecting courses of their choice and on-the-spot classes. This mode not only limits the course resources that students can choose but also limits the sharing of courses (Wang, 2020). At the same time, the teaching platform ignores the individual differences of students and does not give too much consideration to personalized recommendations for students with different interests. Thus, the conventional course selection mode fails to meet the differentiated learning needs of students and also wastes educational resources (Song et al., 2021). Therefore, an appropriate course resource recommendation system can provide more intelligent recommendation services for students. In recent years, with the continuous development of information and communication technologies, electronic education and teaching have been widely promoted and applied. The traditional education system is still the main mechanism of learning, but it has been unable to meet people's demands for personalized resources. In this context, the personalized multimodal multimedia course resource-sharing platform has developed rapidly in recent years (Byun & Lee, 2021).
In the personalized course resource recommendation system, students can obtain more course resource information. The system can also dynamically obtain students' preferences over time, analyze the characteristics of each student, and actively recommend course resources with a high matching degree according to certain standards. Thus, the system can more accurately realize personalized recommendations for courses (Cheng et al., 2021). However, with the continuous increase of resource-sharing platforms, the number of network-shared resources has also increased steadily year by year, which makes the problem of “information overload” more serious (Chaabi et al., 2020; Xu et al., 2021). At the same time, online course platforms often only provide online teaching videos. With the continuous change in students' interests, students cannot communicate with teachers offline and face-to-face, which significantly affects learning (Liu et al., 2020). Therefore, there is an urgent need to develop personalized course recommendation methods in the context of multimodal multimedia cloud platform education to better manage course resources, focus on the practical value, and use the efficiency of college courses in order to meet the different demands of different students in learning. The personalized course recommendation methods will enable students to obtain and understand course resources dynamically and choose more suitable courses.
Course recommendation is a service mode that recommends personalized learning resources to users according to their historical behavior information. From the initial method based on collaborative filtering, to the methods of machine learning and matrix decomposition, and then to the deep learning-based recommendation method, it is all to achieve personalized course recommendation and gradually improve the effect of course recommendation (Zheng 2021). The traditional recommendation methods cannot effectively use the auxiliary information and its potential relationship, resulting in a cold start, sparse data, and other problems. With the growth of data volume and the improvement of computing power, the combination of deep learning and course recommendation provides new ideas for solving the above problems (Yang & Gao, 2022; Nitu et al., 2021). The research of deep learning methods for personalized course recommendation has become the focus of current education research.
Therefore, given the lack of personalization and unreasonable utilization of resources in most of the existing course recommendation systems, this paper proposes a personalized course resource recommendation method using deep learning in the online multimodal multimedia education cloud platform. The innovation of the proposed method lies in: