Article Preview
Top1. Introduction
The form of “Internet+Education” has expanded the scale of online education users, while users’ recognition and acceptance of online education are also continuously improving (Apaza et al., 2014; Khalid et al., 2022). However, there are numerous courses available online offering an increasing number of options to choose from (Apaza et al., 2014; Jing & Tang, 2017). With so many options available, it is extremely difficult for users to choose high-quality and suitable courses. Therefore, personalized course recommendations came into being (Obeidat et al., 2019; Mondal, 2020; Yang & Cai, 2022; Cuesta, 2010).
With the explosive growth of the number of courses, tens of thousands of online course resources provide convenience for learners, but also easily confuse users in their choices, resulting in “information confusion” (Yang et al., 2014; Xiao et al., 2018). Therefore, it is important to build an efficient course recommendation method to provide personalized recommendations for users. Recommendation algorithms have gained significant attention as an effective method to solve the “information maze” problem in academia and industry and have been widely used (Garg & Tiwari, 2016; Yang et al., 2014; Wu et al., 2020; Lin et al., 2021). Fire recommendation algorithms attempt to mine a collection of items that meet users’ interests from massive data. Currently, recommendation algorithms have achieved significant success in many fields including movies, music, news, medical care, and education (Li et al., 2018; Zhou et al., 2022). However, compared to these scenarios, Massive Open Online Courses (MOOC) resource recommendations often face more sparse interactive data and more complex backgrounds. On the one hand, learners typically do not choose too many courses because learning a course requires a lot of time and effort. According to statistical data, each course typically lasts for several weeks, while the average number of course registrations for learners is less than two (Hou et al., 2018; Vlasenko et al., 2022; Zhang et al., 2019). On the other hand, the open nature of MOOC allows it to accommodate many similar courses, which may have diverse approaches to conveying knowledge, ultimately leading to user confusion when selecting a course and posing a serious challenge for generating personalized learning resource recommendations. The course recommendation algorithm recommends courses that may be of interest to users by studying their interest in course selection, historical course selection behaviour, and course attributes. This approach effectively alleviates the problem of information overload and improves the efficiency of course selection and the online experience of users (Wang, 2022; Wang et al., 2020). The key to course recommendation lies in accurately positioning each user's learning goals and needs and finding the most suitable course for users (Gao et al., 2022).