PredictionofStudents’ Academic Performance UsingData Mining Techniques

Abstract

Data mining is the automatic search of huge data to discover patterns and trends that go beyond simple analysis. The high rate of student’s failure is one of the major problems and represents a worry for many universities because no rule for distribution specialization, it's difficult to determine which specialization is better for the student. This study proposed a model for Predicting the best specialization for the student. Weka data mining tool used to evaluate performance of student’s. The data consists of academic information contain 1402 records from University of Kordofan, Faculty of Economics. The experiment conducted using three algorithms Naive Bayes classifier, j48 and Random Forest to predict best student's specialization. Then, Apriori algorithm was also applied to find close correlation between courses and specialization. Results showed that the best technique is J48 classifier was achieved 97.6% of accuracy which is better than Naïve Bayes and Random. The Apriori algorithm used for generates strong rules that helped to identify if the student’s academic qualify to study the Specialization and relationship between specialization and courses. The experiment conducted generated strong rules their number are fifteen with (Support 11% and Confidence 100%).