Performance Evaluation and Comparison of Classification Algorithms for Students at Qassim University

The level of the student may change at the undergraduate level from his level in intermediate or secondary school. Educational and personal factors may affect the performance of university students, especially in the first year of university. In this study, we measure a relationship between a Grade Point Average (GPA) of the students at Qassim University in the first year, and variety of personal and educational factors of the students, such as parents education, number of hours using smartphones and the degree of Scholastic Achievement Admission Test. To get the results, we used six different classification algorithms to measure accuratly for attributes between these algorithms and find the best algorithm that gives us a high accuracy percentage for prediction of students performance. We apply these algorithms using RapidMiner platform, which is one of the most famous and strong data mining tools. Based on the overall comparison of these six classification algorithms, the Neural Network algorithm had the highest accuracy percentage 84 % and the lowest was 56.55 % of the AdaBoot algorithm. Then, we apply association rule to check if some rules demonstrate the existence of direct relationships between personal and educational factors and student performance or no.