Prediksi Kelulusan Peserta Seleksi Nasional Berdasarkan Prestasi (SNBP) Menggunakan Pendekatan Ensemble Learning

Millaturrosyidah, Yusna (2025) Prediksi Kelulusan Peserta Seleksi Nasional Berdasarkan Prestasi (SNBP) Menggunakan Pendekatan Ensemble Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5025211254-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Pendidikan berkualitas merupakan kunci utama dalam mencapai tujuan keempat Sustainable Development Goals (SDGs), yaitu pendidikan yang inklusif dan bermutu. Hal ini mendorong banyak siswa untuk melanjutkan pendidikan ke perguruan tinggi melalui jalur Seleksi Nasional Berdasarkan Prestasi (SNBP) sebagai salah satu jalur masuk Perguruan Tinggi Negeri (PTN). Namun, tingginya tingkat persaingan pada jalur SNBP membuat siswa dan institusi pendidikan memerlukan sistem prediksi yang dapat memperkirakan kelulusan calon mahasiswa berdasarkan nilai akademik dan prestasi. Penelitian ini mengembangkan model prediksi kelulusan siswa yang mendaftar melalui jalur SNBP dengan menggunakan pendekatan ensemble learning. Model ensemble learning yang digunakan antara lain Random Forest, XGBoost, LightGBM, CatBoost, dan Soft Majority Voting. Untuk menangani data yang tidak seimbang, digunakan metode CTAB-GAN dalam pembuatan data sintetis, serta dilakukan penambahan fitur pembobotan prestasi siswa dan nilai mata pelajaran pendukung. Hasil penelitian menunjukkan performa terbaik pada pendekatan ensemble learning menggunakan model CatBoost dengan accuracy 0,9486, precision 0,9492, recall 0,9486, dan F1-Score 0,9488. Setelah dilakukan hyperparameter tuning, performa model CatBoost mengalami peningkatan dengan accuracy 0,9504, precision 0,9517, recall 0,9504, dan F1-Score 0,9508. Sistem prediksi yang dikembangkan dapat digunakan sebagai bahan pertimbangan bagi institusi dalam proses seleksi penerimaan mahasiswa baru, sekaligus memberikan gambaran prediksi kelulusan bagi calon pendaftar SNBP.
=================================================================================================================================
Quality education is a key element in achieving the fourth goal of the Sustainable Development Goals (SDGs), which is inclusive and equitable quality education. This motivates many students to pursue higher education through the National Selection Based on Achievement (SNBP), one of the admission pathways to State Universities (PTN). However, the high level of competition in the SNBP pathway requires students and educational institutions to have a prediction system that can estimate the admission of prospective students based on academic scores and achievements. This research conducts prediction of student admission who apply through the SNBP pathway using an ensemble learning approach. The ensemble learning models used include Random Forest, XGBoost, LightGBM, CatBoost, and Soft Majority Voting. To handle imbalanced data, the CTAB-GAN method is employed to generate synthetic data, along with the addition of feature engineering for student achievement weighting and supporting subject scores.. The research results show the best performance using the ensemble learning approach with the CatBoost model, achieving an accuracy of 0,9486, precision of 0,9492, recall of 0,9486, and F1-Score of 0,9488. After hyperparameter tuning, the CatBoost model performance improved with accuracy of 0,9504, precision of 0,9517, recall of 0,9504, and F1-Score of 0,9508. The developed prediction system can be used as consideration material for institutions in the new student admission selection process, while providing admission prediction insights for SNBP applicants.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Ensemble learning, Prediksi Kelulusan, SNBP, Admission Prediction, Ensemble learning, SNBP
Subjects:	T Technology > T Technology (General)
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Informatics Engineering > 55201-(S1) Undergraduate Thesis
Depositing User:	Yusna Millaturrosyidah
Date Deposited:	28 Jul 2025 09:13
Last Modified:	28 Jul 2025 09:13
URI:	http://repository.its.ac.id/id/eprint/122242

Actions (login required)

View Item