Using data mining to discover the correlation between web learning portfolios and achievements

Internet learning is different from traditional classroom teaching. This is because there is no actual contact between teachers and students; therefore, it is difficult for teachers to keep track of studentspsila learning conditions. By analyzing the learning portfolios of students studying through online learning, teachers can observe their learning activities. This study conducted a cluster analysis on each attribute of studentspsila learning portfolios. By observing the activities of individuals in each cluster and investigating their activities at every stage, the behavioral styles were determined and the appropriate warning messages were sent to students in certain clusters. The learning portfolio analysis system proposed in this study provides statistical and data mining techniques, focusing on studentspsila assignment grades, test results, and records of online learning portfolios, for explorative analysis. It also provides a remedial mechanism for teachers to send out adaptive warning messages according to studentspsila learning portfolios. The evaluation in this study demonstrates that the sent warning messages can affect studentspsila learning behaviors and achievements.