Profiling students on their course-taking patterns in higher educational institutions (HEIs)

Data mining techniques have been found useful in understanding and enhancing student performance as well as decision making related to teaching and learning in HEIs. Literature review enabled the choice of time to degree and cumulative grade point average (CGPA) as examples of student performance factors for investigation. Student features that could be extracted using SQL query from student dataset and related to time to degree and CGPA were analysed. However hidden features were extracted using data mining only. Course taking pattern was used as an example of the hidden feature of students to determine time to degree and CGPA. Profiling of students was considered as an important decision that affects teaching and learning. Clustering was used as the data mining model to profile students whose performance in terms of time to degree and CGPA was related to course taking pattern of students and other features. K-means and EM algorithms were used to generate clusters of student profiles. K-means algorithm did not produce meaningful clusters whereas EM algorithm produced 10 clusters that could be interpreted. Results showed that the best cluster was rated as excellent and only one such cluster was found and it was found that there are other features other than course taking pattern that could determine the profiling of students which could be used in determining time to degree and CGPA and providing better academic support to students.