Student Graduation Predictions Using Comparison of C5.0 Algorithm With Linear Regression

Abstract

Technological advances supported by human knowledge have a very good influence on data and information storage technology, including in predicting student graduation (Graduation Prediction) on time, by applying several existing algorithms. In this study, researchers used the C5.0 Algorithm and Linear Regression. The concept of the research is to compare two algorithms, namely C5.0 and Linear Regression to the case of graduating students on time. Based on the length of study, students who graduated correctly amounted to 651 (91%) with a male gender of 427 students and a female gender of 224 students while those who did not pass (late) correctly amounted to 64 (9%) with a male gender totaling 53 students and female gender totaling 11 students from 2017-2020. Comparison results The R2 score from the C5.0 algorithm reached 96.85% (training) and 93.