RESEARCH ON THE DESIGN OF INTELLIGENT EDUCATION SYSTEM BASED ON BAYESIAN TRACKING MODEL AND DEEP LEARNING KNOWLEDGE TRACKING MODEL

Abstract

Smart education allows students to learn at any time regardless of time and place, and with the combination of advanced technologies such as big data and artificial intelligence, valuable information can be mined from the massive amount of student history data, thus facilitating personalized learning and teaching for students and teachers. Knowledge tracking can get the mastery level of students' knowledge based on their historical behaviors, which can provide students with a good learning feedback, which is also an issue of great importance in smart education. Only when the current learning situation of students is clear, the next learning plan can be better determined, enabling students to understand their learning status and teachers to tailor their teaching to their needs. In this thesis, we propose a new knowledge tracking model based on parametric iterative inference, which takes into account the relationship between related knowledge points under the same topic and designs a multi-objective optimization loss function to improve the consistency of inference. Through experiments on several publicly available datasets and comparisons with several tracking models, including classical and recent ones, the prediction performance of the model is not only better, but also particularly outstanding in terms of consistency. The system is divided into a teacher side and a student side. The student side mainly provides online course viewing, exams, wrong question records, knowledge mastery visualization, and the teacher side provides course management and test thesis management.