Downloads: 166 | Weekly Hits: ⮙1 | Monthly Hits: ⮙1

Research Paper | Computer Science & Engineering | Egypt | Volume 6 Issue 10, October 2017

Early Prediction of Student Success Using a Data Mining Classification Technique

Mohamed Hegazy Mohamed | Hoda Mohamed Waguih

Abstract: Nowadays the amount of data stored in educational databases is increasing rapidly. These databases contain hidden information which can be used for the improvement of student academic performance in higher education systems. Predicting student's academic performance beforehand can help management, faculty as well students to make timely decisions. A data mining technique is used to study the data available in the educational field, and bring out hidden knowledge for decision making. Data mining classification technique was used through three decision tree methods, namely J48, Rep tree, and Random Tree. This paper is an attempt to apply the data mining techniques, particularly classification, to help students in enhancing their results and the quality of the higher educational system by early prediction of student success using decision tree methods. In this study, the main attributes that affect the student performance were determined to predict students? final grade early. For this purpose, we have used real data obtained from the managerial higher institute ?Tammoh? in Giza-Egypt for first year students. Weka data mining tool was used to generate and compare the classifier models of the selected algorithms and results were reported. A ranker search method was then applied to rank the best five attributes in data by using Info Gain ranker to filter the most important rules of the selected classifier model. Final results show that applying J48 algorithm with ranker search method can enhance the generated rules and help management predict early weak students and take appropriate decisions to prevent them from failure and thereby enhance students' academic performance.

Keywords: Data Mining, Classification, Decision Trees, Higher Education

Edition: Volume 6 Issue 10, October 2017,

Pages: 126 - 131

Similar Articles with Keyword 'Data Mining'