MODELING STUDENT GRADUATION IN PTN ENTRANCE SELECTION THROUGH REPORT CARD SCORES USING STOCHASTIC GRADIENT DESCENT CLASSIFICATION

Abstract

There are many ways to enter state universities in Indonesia, namely by taking an entrance test selection or by selecting a report card. The simplest way is to take part in the section based on report card scores because students are not tired of preparing to take part in the selection. Still, in the Mojokerto area, guidance and counseling teachers do not know whether students' report card scores meet the requirements or do not meet the criteria to take part in the selection, and student selection is still being carried out. Manually after that, the teacher provides a letter of recommendation to take part in the section of students who are deemed to meet the criteria. In this experiment, we tried to build a prediction machine based on existing data, so it is hoped that this machine can provide an overview for consideration for guidance and counseling teachers to give recommendations to students who meet the criteria. From this experiment, a training set with 2764 data was studied using the Stochastic Gradient Descent Algorithm and gave a recall score = 0.178, precision = 0.394, and F1 score = 0.245.