编程学习中学生行为分析与结果预测

Behavior Analysis and Result Prediction of Students in Programming Learning

通过收集RoboMise这一自适应编程学习系统中学生的行为数据,应用统计分析方法进行学习行为分析;基于学习行为数据,利用机器学习方法,对学生解决任务的结果进行预测。研究结果表明:可以根据学习行为对学习结果作出预测,随机森林算法的综合预测效果最优。

This study collects the behavior data generated by students in RoboMise,an adaptive learning system for introductory programming and use statistical analysis methods to analyze learning behavior.Further,based on the learning behavior data,the results of students’task solving are predicted by using machine learning.The results show that the learning results can be predicted according to the learning behavior,and the comprehensive prediction effect of random forest algorithm is the best.