【摘要】 在线课程学习过程中会产生大量的学习行为数据。通过对这些数据进行挖掘分析，能够预测在线课程成绩。本文以安徽师范大学“面向对象程序设计”课程后台数据库作为实验数据集，评估了ID3、C4.5和CART等算法构建的决策树预测模型，其中CART算法最优。研究发现，期中测试分数、1-4章测试分数、章节学习次数和互动得分等指标对期末成绩的影响最大。为了提升模型性能，改进了CART算法。实验结果表明，其能够提升预测准确率，具有较好的应用价值。更多还原

【Abstract】 A large number of learning behavior data will be generated in the process of online course learning.The scores of students’ online courses can be predicted by mining and analyzing these data. Taking the background database of Object Oriented Programming course of Anhui Normal University as the experimental data set, the decision tree prediction models constructed by algorithms ID3, C4.5 and CART are evaluated, in which CART algorithm is the best. The study found that the mid-term test scores, test scores of chapter 1-4,chapter learning times and interaction scores had the greatest impact on the final grade. In order to improve the performance of the model, the CART algorithm is improved. The experimental results show that it can improve the prediction accuracy and has good application value.更多还原