Abstract:

The main goal of educational systems is not only to provide quality of education but also make sure that the students are graduating with good grades. The major challenges of higher education being decrease in students' success rate and their leaving a course without completion. An early prediction of students’ failure may help identifying students who need special attention to reduce fail ratio and to take appropriate action. Therefore a new framework called Faculty Support System (FSS) is implemented using different classification techniques to predict student performance based on specific students attributes such as programing assignments, quizzes, in class group projects, attendance and exams.

The Faculty Support System (FSS) shows that our modeling approach with Naïve Bayes classifier (as pre-classifier tool) can effectively predict student’s final grade, identify students at risk, and adopt programs and practices that help the identified students to enhance their performance before the end of the semester. Finally a decision tree classification algorithm (as post-classifier tool) is implemented to evaluate the correctness of the adopted process. We achieved an accuracy of 81.80% on a set of defined attributes for a group of students majoring in computer science at University of Houston, Texas. The proposed algorithm and analysis can be generalized to evaluate other attributes in different situations while still maintaining the same competitive accuracy ratio for improving student performance and reducing the dropout rate in any educational institution.

Keywords:

Educational data mining (EDM), classification, naïve bayes classifier, decision trees.