INTERNET ADDICTION AND STUDY EARLY WARNING SYSTEM BASED ON MACHINE LEARNING METHODS

In view of Internet addiction and academic warning for university students, questionnaires including student performance, design demography, individual behavior, social relationships and Internet addiction were used as study data, and we proposed two machine learning methods of students' internet addiction and academic warning research based on BP neural network and random forest. The study provided early warning mechanism for teachers and students, and analyzed the important influencing factors to achieve precise intervention. Through simulation analysis, BP neural network and random forest algorithms were chosen to train internet addiction and academic study model respectively, and the accuracy reached 92.286% and 92.742% respectively. The study provided early warning mechanism for teachers and students, and it has certain academic and application values.