Forecasting Student Graduation with Classification and Regression Tree (CART) Algorithm

Abstract

Forecasting Student Graduation with Classification And Regression Tree (CART) Algorithm aims to predict students' graduation in Computer Science Departement of Pakuan University. The method used is the CART Algorithm where this method works by forming a decision tree that can infer the rules of a particular classification of the decision tree. The results of this study resulted ouput in the form of decision trees and determine the students with predetermined criteria are classified in categories such as pass on time or pass not on time. Assessment of the accuracy of the data classification that has been done using confusion matrix validation technique. Once calculated with the matrix, the accuracy rate reached 77.5% in the training data.