PREDICTIVE ANALYTICS MODEL FOR PREDICTING UNDERGRADUATE STUDENT GRADUATION PERIOD

• Main Author: Bela Hanifa, Dzar
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/43859
• Institution: Institut Teknologi Bandung

Delayed graduation for student could lead to various problems for the student and also for campus administration. In order to anticipate this delay, several research has been conducted to build prediction model for predicting student graduation period. However, most of those research only categorize student graduation period into two categories, on time and delayed. Most of the research also only use intracurricular data, whereas data that are related to extracurricular activity is often neglected. This research aims to build predictive analytics model for predicting student graduation period. Those model will predict student graduation period into three categories, early, on time, and delayed. Student activities, both intracurricular and extracurricular, would be used as the data source. The predictive analytics model then would be implemented on a software that will produce graduation period prediction as its output. This research compares the performance of Random Forest Regression and Multiple Linear Regression algorithm using accuracy and mean squared error metrics. Based on those two metrics, Random Forest Regression is the better algorithm for predicting student graduation period.