Since Accumulated Grad-Point Average (AGPA) is crucial in the professional life of students, it is an interesting and challenging problem to create profiles for those students who are likely to graduate with low AGPA. Identifying this kind of students accurately will enable the university staff to help them improve their ability by providing them with special academic guidance and tutoring. In this paper, using a large and feature rich dataset of marks of high secondary school subjects, we developed a data-mining model to classify the newly-enrolled students into two groups; “weak students” (i.e. students who are likely to graduate with low AGPA) and “normal students” (i.e. students who are likely to graduate with high AGPA). We investigated the suitability of evolving fuzzy clustering methods to predict the ability of students graduating in five disciplines at Sultan Qaboos University in the Sultanate of Oman. A solid test has been conducted to determine the model quality and validity. The experimental results showed a high level of accuracy, ranging from 71%-84%. This accuracy revealed the suitability of evolving fuzzy clustering methods for predicting the students’ AGPA.
|Journal||Australian Educational Computing|
|Publication status||Published - 2015|
- Educational data mining
- Evolving fuzzy clustering
ASJC Scopus subject areas
- Computer Science Applications