Hassan Ahmadi Torshizi - Islamic Azad University- Mashad Branch
Alaleh Rangriz - NooreTouba Virtual University, Tehran

The aim of this paper is to predict the students’ academic performance. It is useful for identifying weak students at an earlier stage. In this study, we used WEKA open source data mining tool to analyze attributes for predicting students’ academic performance. The data set comprised of 154 student records and 21 attributes of students registered between year 2006 and 2009. We applied the data set to four classifiers (Naive Bayes, Kstar, MLP and Random Forest) and obtained the accuracy of predicting the students’ performance into either successful or unsuccessful class. the student's academic performance can be predicted by using past experience knowledge discovered from the existing database. A cross-validation with 10 folds was used to evaluate the prediction accuracy. The result showed that Random Forest classifier scored the higher percentage of prediction F-Measure of 91.6%.

ELearning Enrichment, Data Mining, Learning Performance Prediction