Research And Implementation Of Personalized Course Recommendation System Based On Self-attention Mechanism

Nowadays,mobile Internet has injected vitality into online learning,and the online learning industry has ushered in a boom of development again.Educational websites relying on mobile Internet have appeared in large numbers and tend to have large-scale personalized development.Artificial intelligence and big data technology are more widely used in online learning.In recent years,the number and variety of online learning resources have proliferate,making it extremely difficult for students to find the course resources they want among the numerous resources.Coupled with the interference of some irrelevant factors,it is difficult for learners to find effective courses,which will not only consume their learning time,but also greatly reduce their learning interest and efficiency.In addition,learners’ interests will change with time,and these problems will gradually emerge with the development of online learning.In this thesis,the development status of the online learning industry at home and abroad is studied in detail,the educational theory of online learning and the recommendation technology of the current popular learning system are studied and analyzed in depth,which echoes the actual needs of students and combines professional knowledge and big data technology.Finally,a personalized course recommendation system based on self-attention mechanism is designed and implemented.In the early stage of building the recommendation model,this thesis proposes an interest screening mechanism,which considers user feedback,user viewing time and user viewing frequency to screen noise data such as user misclicks that may affect the recommendation results.Then,by considering the influence of time interval on students ’learning interest,this thesis proposes to combine it with self-attention mechanism to obtain students ’interest features.Then the similar users are found through the user information,and the interest characteristics of the user and its similar users are comprehensively considered.Finally,a personalized course list is generated for students,and more reasonable video courses are recommended for them.Finally,experiments show that the personalized recommendation algorithm based on self-attention mechanism is superior to common recommendation strategies and meets the design requirements.In this thesis,the front-end of the personalized course recommendation system based on self-attention mechanism uses the Bootstrap framework to improve the beauty of the system.In the back-end development of the website,this thesis chooses the Django framework of Python language as the business development framework,and the relational database MySQL is used to store important information in the system.The personalized learning recommendation system based on self-attention mechanism is fully designed and easy to operate by users,and the system has been fully tested and improved to ensure that the system can achieve the expected goal,which can effectively select the course videos that users really need from the massive learning resources to recommend,and improve the efficiency and quality of students’ learning.