Prediksi Peringkat Akreditasi BAN PT Program Studi Sarjana Rumpun Ilmu Komputer Menggunakan Klasifikasi Machine Learning

Abstract

Accreditation ranking is one of the causes and indicators chosen by prospective students when choosing a study program in higher education. From the data collected, only 5% of study programs in the Computer Science group have a Superior accreditation rating and an A accreditation rating in LLDikti Region III Jakarta. So it is necessary to know the factors that influence the accreditation ranking. The machine learning methodology used in this approach is K-Nearest Neighbors (KNN) and from the data obtained there are 6 factors that can be strongly suspected to influence the study program accreditation value. The four machine learning models, namely KNN, Gaussian Naïve Bayes Decision Tree and Logistic Regression, it was found that the KNN machine learning model with 2 input variables had the highest AUC value, namely 84.38%. Meanwhile, from the model simulation run by KNN machine learning, 2 input variables can produce relatively accurate prediction results. And the results of cross validation with 10 folds support the selected machine learning with an accuracy level of 80%. In general, the KNN machine learning model with 2 input variables was able to predict the accreditation rating of Study Programs, especially from the Computer Science Cluster.

Keywords – Accreditation, Area Under Curve (AUC), Department of School, Kfold Cross Validation, Machine Learning.