Data mining in higher education: university student dropout case study

In this paper, we apply different data mining approaches for the purpose of examining and predicting students’ dropouts through their university programs. For the subject of the study we select a total of 1290 records of computer science students Graduated from ALAQSA University between 2005 and 2011. The collected data included student study history and transcript for courses taught in the first two years of computer science major in addition to student GPA , high school average , and class label of (yes ,No) to indicate whether the student graduated from the chosen major or not. In order to classify and predict dropout students, different classifiers have been trained on our data sets including Decision Tree (DT), Naive Bayes (NB). These methods were tested using 10-fold cross validation. The accuracy of DT, and NlB classifiers were 98.14% and 96.86% respectively. The study also includes discovering hidden relationships between student dropout status and enrolment persistence by mining a frequent cases using FP-growth algorithm.