学生日常行为模式挖掘及差异性分析

Mining Students’ Daily Behavior Patterns and Analyzing Discrepancies

针对学生日常行为与学业成绩关系问题展开研究.面向反映学生日常行为的手机上下文数据,提出了一种基于关联规则的行为模式挖掘及差异性计算方法.其特点是使用语义化处理方法将数值型数据转换成具有语义信息的数据,采用Apriori算法挖掘关联规则,通过定量计算特征关联规则集合之间的非相似性系数,区分出不同类型学生行为模式之间的差异,进而得出学生日常行为与学业成绩之间的关系和影响,并在公开数据集上对该方法的有效性进行了实验验证.

In this paper,we studied the relationship between students＇ daily behaviors and their academic performance.According to the smartphones＇ sensor context data that reflects students＇ daily behaviors,we proposed a novel behavior pattern mining and discrepancy calculation method based on the technology of association rules.In this method,firstly the numeric data was converted into the form with semantic information by using semantic process algorithm.Secondly,we utilized the algorithm called Apriori to mine the hidden affinities among the behavior data.Then,we constructed a method to quantitatively calculate the non-similarity coefficients between the sets of feature association rules.Finally,we distinguished the similarity and discrepancies among different types of students＇ behavior patterns and then we could get the relationship between the daily behavior and the academic performance.And the validity of the method was verified experimentally in the open data set.