| Summary: | 碩士 === 元智大學 === 資訊工程學系 === 104 === Learning analytics apply big data techniques in analyzing educational data, including students’ academic performance course taken record, personal information, and learning behavior to diagnose student’s learning and identify at-risk students for early intervention. This study adopts student’s student model of core competencies from competency-based learning analytics system at Yuan Ze University and records of delayed graduation and dropout to build a predictive mechanism. This mechanism estimates student’s risk of delayed graduation and dropout and identifies at-risk students. This mechanism not only provides dichotomous identification of at-risk and non-risk students, but also clusters high and low risk groups among at-risk students and non-risk students.
The evaluative results revealed that the precision, recall, and F values of identification accuracy of at-risk students of delayed graduation are 0.74, 0.68, and 0.7. The precision, recall, and F values of identification accuracy of at-risk students of dropout are 0.87, 0.83, and 0.85. The results also showed that groups at high risks had higher delayed graduation and dropout rate than those at lower risks.
|
|---|
