A Comparative Analysis of Classification Algorithms for Students College Enrollment Approval Using Data Mining
Abdul Hamid M. Ragab, Amin Yousef Mohammad Noaman, Abdullah Saad Al-Malaise ALGhamdi, Ayman I. Madbouly · 2014
In big data universities, there may be several problems related to college admission and enrollment due to the increasing volume of students' data applying for higher education. So that there is a need to apply efficient data mining algorithms for better decision making for students' data classification. In this paper, nine classification algorithms are comparatively tested to find the optimum algorithm for students' dataset classification. The KAUODUS+ data base for the preparatory year students are used as an approval dataset for the experimental purposes. The Weka-knowledge analysis tool which is open source data mining workbench software is used for simulation of practical measurements. The classification technique that has the potential to significantly improve the performance is suggested for use in colleges' admission and enrollment applications. Impact of students GPA and their grades qualified materials with respect to their colleges' admission desire are visually analyzed. Results show that C4.5, PART and Random Forest algorithms give the highest performance and accuracy with lowest errors while IBK-E and IBK-M algorithms give high errors and low accuracy.