Research on the Learners'Performance Prediction Combining Theory Driven and Data Driven
Accurately predicting students'academic achievement can identify at-risk students as early as possible and implement targeted interventions on them.This paper takes 160 students taking the Computer Network and Cloud Computing Course in a university as research subjects,combines the theory-driven and data-driven methods,uses cluster analysis and cross analysis to study the effects of student subjective self-evaluation and online learning behavior data on predicting students'performance.It was found that the theory-driven method identified two types of online learning,naming"Deep Learning"and"Surface Learning",there was a significant difference between the two learning patterns in academic achievements.The data-driven method identified four types of online learning models(active,task,passive and more passive),and there were significant differences in learning behaviors and achievements among the students of these four types of learning patterns.The cross analysis was used to identify seven types of learning patterns and there were significant differences in learning engagement,learning strategy,learning behavior and learning achievements among the seven types of students.
万方数据
维普
