基于改进C4.5的E-learning教学辅助系统的研究与实现  

Research and Implementation of E-learning Teaching Assistant System Based on Improving C4.5

作　　者：谢修娟[1] 李香菊[1] 操凤平 孙丽 

机构地区：[1]东南大学成贤学院计算机工程系,江苏南京210000

出　　处：《佳木斯大学学报（自然科学版）》2018年第1期64-67,82,共5页Journal of Jiamusi University：Natural Science Edition

基　　金：江苏高校哲学社会科学基金资助项目"基于混合式学习和翻转课堂的SPOC模式在高校教学中的应用研究"(2016SJD880186);江苏省现代教育技术研究课题(2016-R-46509)

摘　　要：为提高E-learning(数字化学习)中学生自主学习以及教师管理学习的效率,提出将决策树C4.5算法应用于数字化学习平台中的决策分析,设计了基于决策树的E-learning教学辅助系统,根据学生的在线学习行为预测其学习效果,以尽早发现问题。基于Fayyad边界点判定原理和数学的等价无穷小理论,对传统C4.5算法作出两点改进,在E-learning系统中,运用改进的算法先训练出在线学习行为与学习效果间的决策树模型,导出分类规则,而后进行学习效果预测。实验结果表明,改进后的算法具有比较高的预测准确率,能够为学习者和教学者提供决策支持。E - learning （ digital learning） is a way of online learning, m order to improve me efficiency of students＂ autonomous learning and the teachers＂ management of learning, it puts forward that to use decision tree C4.5 algorithm in E -learning teaching system, and to forecast online learning effect of students＂ course learning, so that we can find the problem as soon as possible. Two improvements are made on the traditional C4.5 algorithm based on Fayyad boundary point for determining principle and the mathematical theory of equivalent infinitesimal. In E -learning system, firstly trains the decision tree model of students＇online learning behavior and learning efficiency by improved algorithm. Secondly, exports classification rules, and then forecasts study effect. The experimental results show that the improved algorithm has higher prediction accuracy, and can provide decision support for learners or teachers.

关 键 词：E-LEARNING系统 决策树C4 5算法 在线学习行为 预测分析 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]