Application of The Support Vector Machine Algorithm for Timely Student Graduation Prediction Based on Streamlit Web at The Faculty of Informatics Engineering Nurul Jadid University

Abstract

Universities must provide good education so that they can produce good graduates.There are many factors that influence student graduation rates, one of the problems faced by an educational institution, especially at universities, whether state or private, is finding predictions of student graduation rates on time.One of the technological advances currently available is a system that can predict whether students will graduate on time or not. One of the machine planning algorithms that can be used is the Support Vector Machine.The results of this research were carried out by predicting the on-time graduation rate of students at Nurul Jadid University, Faculty of Engineering, Informatics Study Program. By using the Support Vector Machine method, this research used testing data of 20% of the data from 612 student data with the same 7 attributes. The data obtained 123 data which resulted in 72 student data being on time, 45 student data being late, 4 student data being correct. time and 2 students' data was late. From the results, the accuracy of the training data was 94%, while the results of the accuracy of the testing data received a score of 95%.  And based on the validity test of the Support Vector Machine algorithm, the presentation results obtained were Accuracy levels of 96%, Recall 98%, and Precision 94% from 123 testing data. Next, the model is deployed using Streamlit. Streamlit is an open source Python-based framework designed to help developers build interactive web-based programs in the fields of data science and machine learning. The accuracy rate is very good, this shows that SVM can be applied to predict student graduation rates.