Research and Implement of Course Early Warning System Based on Teaching Behaviour Data

Abstract

Digital early warning research based on the course performance of students, is constantly advancing. However, there are a few scholars combing one of specific course, taking the early predication by digging and analysing the data of teaching behaviour and provide real-time intervention. This research, making Java program design as entry point design an overall framework for early warning system based on the data of teaching behaviour, and combines the certain algorithm to implementing the early course predication model with relatively higher accuracy by gathering and analysing students' past learning data as well as real-time teaching behaviour's data. Practice indicates that, according to the grade and GPA (grade point average) of pre-course as well as study behaviour during week 4 and week 6, there can be predicted course pass rate of students accurately, especially from week5, which has high accuracy of predicting. Moreover, the model is able to provide early intervention and automatically sending learning materials for some students who get unsatisfied performance basing on the predicating data, to improve the pass rate for students.