Prediction of the Quality of Prospective Student Graduation for Determining New Student Selection Using the C4.5 Decisson Tree Algorithm

Abstract

Historical data of former student contains rich informations. This research begins with a hyphothesis that there must be a correlation between former student data history and their success. By deeply observe the historical background of student, a pattern can be determined. Once a pattern is developed, a prediction of success can be projected to new and current students. This information is helpful for the new student admission strategy. In order to find the pattern a machine learning algorithm will be implemented. C45 works very well for machine learning to generate the rule based on attributes impact to the label. C45 is impelemnted to learn student pattern in the past in order to determine the pattern to predict the future students. A dataset is taken from the last 3 years vocational student graduation dataset. Success is defined as national exam and length of study. A final exam grade of senior high school is consider input to C45 classifier. According to experiments result an accuracy of report score prediction to determine success of national examination score is achieved by C.45 algorithm on success prediction. The test score attribute is taken from presence shows the highest impact to the student success, while attributeachievement as the lowest impact to the output.