Modeling an Automated Student’s Performance Predictor by Using Decision Tree

Abstract

The student performance is essential key factors in the educational field as the goal of our country is to promote the students with high quality of education on the market. Nowadays, the students enroll on the modules or courses during the registration period and they pay the school fees according to registered modules. The students who fail the module must retake it in the next years. All High learning institutions use management information system that holds the information about their students, these information’s are potential to the different task. Data mining   enables them to extract and discover the unknown knowledge   from data stored in the database. This field of data mining uses the academic settings called Educational data mining (EDM). Predicting student performance is one of the applications conducted in this field. In this paper, we developed an automated model that will predict student performance, this enables the HLI to analyze and model the performance of their students at the middle of their study. Classification techniques are one technique used   for predicting useful knowledge through the historical information by data mining tools. We analyzed data by using   three decision tree algorithms named Iterative Dichotomiser3 (ID3), J48 and classification and regression (CART) decision tree. After analyzing data, we find that the ID3 with an accuracy of 70.4% was the best algorithm used because it had higher accuracy compared to J48 and CART