Course Selection Recommendation System Based On Behavior Analysis

The recommendation system is a service and technology that intelligently provides personalized information and advice to customers along with the rise of the Internet and big data era.It analyzes the user's preferences and traits to achieve personalized recommendations for users and even precise services.Currently,the recommendation system has a relatively successful application in the fields of e-commerce,video and audio,and social friends.In order to comply with the development trend of social diversification,modern higher education is also undergoing continuous reforms,offering a wide variety of elective courses,and students are increasingly diversified in choice,which also causes students to face the course of course selection.The problem of information overload.In addition,in the actual course selection process,most students are not clear about their interests,and they are not very familiar with their professional employment situation,with certain blindness.Therefore,the demand for the elective recommendation system has become direct and urgent.At present,most college college selection strategies are based on students'academic achievement and historical elective records,and recommend students' courses.This method lacks analysis of other individual data of students and excavation of course characteristics and course selection process,resulting in the final recommendation result is not Very satisfactory.Based on the current mainstream course recommendation system,this paper proposes an algorithm framework based on student behavior analysis.By introducing a large number of students'learning behavior data,the students'personality characteristics are extracted,and the social relationship between students and students is calculated.Complete behavioral portraits,combined with training programs and curriculum,provide students with personalized curriculum recommendations.286 undergraduates from the 2016 Chemistry College of University were selected as test sets.The measured data show that the recommendation algorithm based on student learning behavior in this paper has improved the performance of the related evaluation indicators such as accuracy and recall rate compared with the traditional algorithm,and can basically realize the personalized recommendation of the course.At the same time,the performance on the cold start of the new students is also satisfactory.