Personalized Course Recommendation System Based On XDeepFM In MOOC Environment

With the advancement of the national education information education strategy,major online education institutions have sprung up,major colleges and universities have also launched online courses.However,the massive and abundant learning resources in online platforms also bring about problems such as "information trek" and "information overload"for learners.How to accurately recommend personalized learning resources for target users has become an urgent problem for the platform to solve.Therefore,the application of personalized recommendation technology in the field of online education is a feasible solution.It solves the problem that users are difficult to search for the learning resources they are interested in from the mass of data,and it can also increase users 'product love and learning motivation.Deep learning has been widely used because of its efficient performance in image recognition,speech recognition and other fields.It also proposes effective solutions to solve the problems of sparseness,difficult to extract complex features and cross-mining features in traditional recommendation systems.To this end,this article proposes a personalized course recommendation system based on XDeepFM in the MOOC online platform of Chinese universities.The main research work is as follows:1.An overview.of the current research progress of deep learning-based recommendation and educational learning resources,the ideas,principles,specific processes and their advantages and disadvantages of commonly used recommendation algorithms are described,and the basic model of deep learning is briefly introduced,and finally proposed This article evaluates the experimental methods and evaluation indicators of personalized recommendation systems.2.A personalized course recommendation model based on the Bert network model integrated with natural language processing under the XDeepFM framework was established.First,through the course summary,course review information,and the history of the user's learned courses,the user's potential favorite preferences are obtained.The Bert network model is used to convert text information into feature values as supplementary inputs to the model.Then based on the XDeepFM recommendation framework,the proposed model is used for parameter setting and training,and the root mean square error(RMSE)is selected as the evaluation index.At the same time,it compares with the performance of other recommended algorithm models on the MOOC dataset of Chinese universities.The experimental results show that the proposed model is effective.The information features of courses and users are used to explore the potential interests of users and improve the accuracy of recommendations.3.According to the practical application scenario of Chinese university MOOC,design the functions of user login and registration,course management,user information management,recommended courses,etc.,and develop the recommendation system based on the system architecture to realize all the functions designed by the recommendation system,and finally the personality The recommendation system has been systematically tested to verify the correctness of the function and stability of the performance.