Research on Mining and Detection Method of Abnormal Learning Behavior
Yan Cheng, Yongchun Miao, Ping-Fei Tan, Yanwen Qu · 2016
In order to mine the abnormal learning behavior of learner in virtual learning community and carry on personalized supervision and guidance, behavior filtering model based on the factor analysis of the behavior is constructed to solve the problem of the relationship between behavior factors in the learning behavior vector space model. In view of the disadvantage of neglecting local abnormal points in a global mining, a community abnormal behavior detection algorithm based on the clustering analysis method was proposed. The learners who have the similar behavior patterns were gathered into a cluster using clustering method before the abnormal behavior in each cluster was detected, which improved the accuracy of the results and reduce the complexity of computational. Finally, using our learning platform of Moodle for the experimental platform from which data was collected and analyzed, the experimental results verified that the algorithm has high accuracy and good practicability.