PERFORMANCE TEST OF THE K-NEAREST NEIGHBORS (KNN) ALGORITHM ON POLYMEDIA NEW STUDENT ADMISSION PERFORMANCE PREDICTION
Dwi Riyono, Cholid Mawardi, Freniko AH

Politeknik Negeri Media Kreatif

Abstract

New Student Admission (NSA) for tertiary institutions is the main and initial gate in academic implementation operations in the tertiary environment. NSA is key for universities in managing students so that the lecture process runs well. NSA at the Creative Media State Polytechnic experienced a downward trend when the admission system was merged into a single process between the Diploma vocational program and PTN Academic. This is a big challenge for Polimedia amidst the link & match policy between vocational universities and industrial partners. As well as the need to cover NSA Polimedia^s capacity, which in 2023 will only reach 72%. A special strategy is needed to increase the capacity figures for Polimedia^s NSA, of course with the help of computing with Machine Learning models in making predictions to get performance values for the following year^s NSA in 2024. From the results of these predicted values, it is hoped that experiments and model evaluations will be obtained, by randomly dividing the training data and test data with several experiments to obtain the highest accuracy values. The results of the highest accuracy in predicting performance will help Polimedia in measuring promotional strategies for the following year.

Keywords: Keywords: NSA, KNN, prediction, performance, Machine Learning

Topic: Digital Media