高校学生上网行为分析与数据挖掘研究

发布时间：2018-07-14 19:14

【摘要】：网络对于学生的学习既有正面作用,又有负面作用。为了制定合理、有效的网络管理措施,本研究以学生上网数据和学生成绩数据为研究对象,采用决策树、关联规则、逻辑回归三种数据挖掘方法对学生上网行为相关属性与学生学习质量之间的关系进行了研究。研究结果表明,学生上网行为的相关属性如"上网时长""入流量""出流量""总流量""上网次数"等均对学生的学习质量有不同程度的影响,且均呈现负相关关系。其中,"上网时长"是影响学习质量的主要因素,可以根据数据挖掘发现的关键数据节点,制定相应的网络管理措施限制学生过度上网。三种数据挖掘方法都较好地实现了预期效果,得出的结论基本一致,其中关联规则挖掘算法的总体性能最好,决策树算法的总体性能略好于逻辑回归算法。
[Abstract]:The network has both positive and negative effects on students' learning. In order to make reasonable and effective network management measures, this study takes students' online data and student achievement data as research objects, adopts decision tree and association rules. Three data mining methods based on logical regression are used to study the relationship between students' online behavior related attributes and students' learning quality. The results show that the related attributes of students' online behavior, such as "duration of Internet access", "flow in and out", "total flow", "Internet times" and so on, have different effects on students' learning quality, and all of them have a negative correlation. Among them, "time on the net" is the main factor that affects the quality of learning. According to the key data nodes discovered by data mining, the corresponding network management measures can be formulated to limit students' excessive access to the Internet. The three data mining methods have achieved the expected results, and the conclusions are basically the same. Among them, the overall performance of association rule mining algorithm is the best, and the overall performance of decision tree algorithm is slightly better than that of logical regression algorithm.
【作者单位】：南通大学研究生院;南通大学交通学院;
【基金】：全国教育信息技术研究“十二五”规划立项重点课题“基于数字化校园的学生上网行为数据挖掘技术研究”(项目编号:136221504) 江苏省研究生教育教学改革研究与实践重点课题“基于MOOC的研究生教学模式改革研究与实践”(编号:JGZZ16_072) 江苏省现代教育技术研究2014年度重点课题“基于E-Learning的高校教育大数据挖掘分析与研究”(课题编号:2014-R-30418) 南通大学自然科学类研究项目“基于主动推送的高校学生信息平台关键技术研究”(编号:13Z034)的阶段性研究成果
【分类号】：TP311.13;G645.5

【参考文献】